Prediction of the occurrence of calcium oxalate kidney stones based on clinical and gut microbiota characteristics

Abstract

Purpose To predict the occurrence of calcium oxalate kidney stones based on clinical and gut microbiota characteristics. Methods Gut microbiota and clinical data from 180 subjects (120 for training set and 60 for validation) attending the West China Hospital (WCH) were collected between June 2018 and January 2021. Based on the gut microbiota and clinical data from 120 subjects (66 non-kidney stone individuals and 54 kidney stone patients), we evaluated eight machine learning methods to predict the occurrence of calcium oxalate kidney stones. Results With fivefold cross-validation, the random forest method produced the best area under the curve (AUC) of 0.94. We further applied random forest to an independent validation dataset with 60 samples (34 non-kidney stone individuals and 26 kidney stone patients), which yielded an AUC of 0.88. Conclusion Our results demonstrated that clinical data combined with gut microbiota characteristics may help predict the occurrence of kidney stones.

Theranostic roles of machine learning in clinical management of kidney stone disease

“…Nevertheless, only few studies have used clinical characteristics to assist KSD diagnostics. Using clinical and gut microbiota traits, one can predict the development of CaOx KSD [58] . Recently, Kavoussi et al [59] have used 24-h urine and clinical data to predict urinary abnormalities.…”

“… Study/Reference Year Objective Input Method(s) Accuracy (%) Sensitivity (%) Specificity (%) AUC Längkvist et al [52] 2018 Detecting kidney stone in CT images CT images Deep learning (CNN) n/a 100.00 n/a 0.997 Parakh et al [20] 2019 Detecting ureteral stone in CT images CT images Deep learning (CNN) 95.00 94.00 96.00 0.954 De Perrot et al [48] 2019 Differentiating kidney stones and phleboliths in low-dose CT (LDCT) images Radiomics features extracted form LDCT Machine learning (AdaBoost) 85.10 91.70 78.30 0.902 Cui et al [55] 2021 Detecting and scoring kidney stone score based on S.T.O.N.E. nephrolithometry Non-contrast CT (NCCT) images Deep learning (CNN) & Machine learning (3D U-Nets) n/a 95.90 n/a n/a Sudharson et al [53] 2021 Detecting kidney abnormalities from noisy ultrasound images Ultrasound images Machine learning (SVM) & Deep learning (CNN) 87.31 at noise level = 0.02 n/a n/a n/a Yildirim et al [54] 2021 Detecting kidney stone using coronal CT images CT images Machine learning (XResNet50) 96.82 95.76 97.00 n/a Xiang et al [58] 2021 Predicting calcium oxalate kidney stone Patients and microbiota characteristics Machine learning (RF) n/a n/a n/a …”

Gut microbiota in patients with kidney stones: a systematic review and meta-analysis

“…In addition to 16S rRNA sequencing, two studies combined shotgun sequencing (29), frc-gene amplicon sequencing and denaturing gradient gel electrophoresis (DGGE) ngerprinting(32) to characterize microorganisms. Three studies (27,28,33) failed to report α and β diversity. The intestinal ora with statistical differences reported in each study was inconsistent and listed in Table 1.…”

“…According to Tang et al(26), P. aeruginosa (AUC = 0.947) and Escherichia coli (AUC = 0.840) could be used to accurately categorize patients with nephrolithiasis. Xiang et al (33) reported that the relative abundances of genera Flavobacterium, Rhodobacter, Gordonia were found useful in predicting kidney stones (AUCs ranging from 0.682 to 0.763). They found that using data from the three genera and four clinical indicators (oxalate concentration, acetic acid concentration, citrate concentration, phosphorus concentration) together produced predictions that were more accurate than those made using just general or clinical data, and random forest developed the most accurate model (AUC = 0.936).…”

Role of microbiome in kidney stone disease

“…A recent first systematic review of gut microbiota and KSD has shown that α-diversity decrease of the intestinal microbiota in KSD patients is very common, together with significant differences in β-diversity compared to healthy controls [40 ▪▪ ], therefore there is sufficient evidence to support that the Intestinal microbiota in patients with nephrolithiasis is altered at a functional and compositional level. In fact, a study has been carried out so that, using machine learning tools, the presence of calcium oxalate stones can be predicted based on the patients clinical and gut microbiota characteristics [41 ▪ ].…”