Urinary tract infections (UTIs) are among the most common community-acquired and hospital-acquired infections worldwide. Urine culture is considered the gold standard for laboratory diagnosis of UTI (Del Ben et al., 2023; Monsen and Ryden, 2017; Herráez et al., 2018; Erdman et al., 2017). However, the culture process typically requires 24–48 h and represents a significant workload for microbiology laboratories. In clinical laboratories, more than half of the samples sent for urine culture are negative (Del Ben et al., 2023; Monsen and Ryden, 2017). Many patients are also started on empirical antibiotics before a diagnosis is made, contributing to the development of antibiotic resistance (Del Ben et al., 2023; Monsen and Ryden, 2017; Herráez et al., 2018; Erdman et al., 2017; Kim et al., 2019).

The development of screening methods with high sensitivity and negative predictive value (NPV) to identify negative samples before culture is essential. A rapid and reliable screening method would shorten the time required to report negative results, reduce the laboratory workload and costs, and facilitate earlier clinical decision-making, including patient discharge. It would also reduce unnecessary empirical antibiotic use and, consequently, the development of antibiotic resistance (Del Ben et al., 2023; Monsen and Ryden, 2017; Herráez et al., 2018; Erdman et al., 2017; Ilki et al., 2017; Wang et al., 2023; Millán-Lou et al., 2018).

Urinalysis is essential for diagnosing UTIs, and recent advances in automated analysis technologies, particularly flow cytometry-based systems, have enabled comparisons with conventional culture methods. Most studies have mainly focused on evaluating leukocytes and bacteria together (Del Ben et al., 2023; Monsen and Ryden, 2017; Chun et al., 2022).

In addition to conventional parameters, emerging urine analyzers have introduced novel particle classification capabilities, including differentiation between bacterial morphotypes and other formed elements. These advances may offer incremental diagnostic value beyond traditional leukocyte and total bacteria counts, particularly in distinguishing true infection from contamination. However, the clinical utility and added value of these extended parameters remain insufficiently validated in routine practice. Therefore, further evidence is needed to clarify whether these advanced analytical outputs can meaningfully improve screening performance and optimize culture utilization in real-world laboratory settings.

In this study, midstream urine samples sent to the microbiology laboratory for culture were simultaneously analyzed using the MUS-3600 (Dirui Industrial Changchun, China). The diagnostic performance of the analyzer was evaluated by comparing its results with those of urine culture as the reference standard.

When culture results were evaluated in this study, 153 (20.2%) of 759 samples showed growth of specific uropathogens, while 606 (79.8%) showed no growth. Positive samples were predominantly from female patients (111, 72.5%).

Table 1 shows the distribution of demographic and clinical findings based on the presence or absence of culture growth. A statistically significant difference was observed between the groups for all demographic and clinical variables except hyaline cast, crystal, calcium oxalate, and pH (p > 0.05). Among the 41 nitrite-positive samples, 39 showed growth in culture.

When examining the distribution of microorganisms, Escherichia coli was the most dominant uropathogen, accounting for 44.4% of the cases. It was followed by Streptococcus agalactiae (15.7%), Enterococcus faecalis (14.4%), and Klebsiella pneumoniae (11.8%) (Table 2).

Table 3 presents the results of correlation analyses evaluating the relationship between WBC, bacilli count and other demographic and clinical variables. WBC showed a very weak but statistically significant correlation with hyaline cast, granular cast, yeast, crystal, mucus; a weak significant correlation with nitrite and suspected cocci; a moderate significant correlation with WBC cluster, squamous epithelial cell, non-squamous epithelial cells, transitional epithelium and total bacterial count; and a strong positive significant correlation with strip leukocytes.

Bacillus showed a very weak positive significant correlation with yeast, pH; a weak with non-squamous epithelial cell, WBC cluster, leukocyte, transitional epithelial cell, moderate positive significant correlation with squamous epithelial cell, suspected cocci, nitrite, and total bacterial count. A very weak negative correlation with age was also observed.

The ROC curves for WBC, bacillus, suspected cocci, and total bacterial count, based on culture growth in male and female patients, are shown in Figure 1.

The diagnostic performance metrics at different cut-off points, along with gender-specific analyses, are presented in Table 4. The area under the curve (AUC) for WBC was 0.798, and the cut-off value was determined to be 39, with a sensitivity of 63.4%, specificity of 84.5%, PPV of 50.8%, NPV of 90.1%, and an elimination rate of 67.5%. For bacilli and suspected cocci, the AUC values were 0.778 and 0.642, respectively, with cutoff values of 103 and 859. The sensitivity, specificity, PPV, NPV, and elimination rate for bacilli were 62.1%, 88.1%, 56.9%, 90.2%, and 70.4%, respectively, while for suspected cocci, these values were 44.4%, 78.7%, 34.5%, 84.9%, and 62.8%, respectively.

When WBC (>39) and bacillus (>103) were evaluated together as a screening method (both parameters positive), the highest overall specificity (96.7%), PPV (78.3%), NPV (87.9%), and elimination rate (77.2%) were observed, although sensitivity decreased to 47.1%. In subgroup analysis, this combination gave an elimination rate of 86.9% (NPV: 93.9%) in males and 68.9% (NPV: 82.1%) in females. The performance was notably better in male patients, with WBC alone showing a sensitivity of 90.5% and NPV of 98.5% at the same cut-off.

Urine samples constitute a substantial proportion of laboratory workload, and a large proportion are reported as either culture-negative or contaminated. In this context, rapid and reliable screening strategies are essential to reduce unnecessary culture requests and inappropriate empirical antimicrobial use (Ilki et al., 2017; Dong et al., 2022; Íñigo et al., 2016; De Rosa et al., 2018).

In the present study, we evaluated the performance of the MUS-3600 urine analyzer as a screening tool by comparing its parameters with urine culture results. The strong concordance observed between nitrite positivity and culture growth supports the established role of nitrite as a reliable rule-in indicator in routine practice. In addition, the predominance of Escherichia coli among uropathogens is consistent with well-established epidemiological patterns reported in previous studies (Monsen and Ryden, 2017; Erdman et al., 2017; Kim et al., 2019).

Among the evaluated parameters, WBC demonstrated moderate diagnostic accuracy with a relatively high negative predictive value, supporting its utility as a screening marker for excluding infection. Feng Dong et al. reported a lower NPV for WBC (75.3%), whereas the higher NPV observed in our study may be attributable to differences in sample size and study design (Dong et al., 2022).

Bacillus showed a comparable level of diagnostic performance to WBC, whereas suspected coccus exhibited clearly inferior performance, indicating limited clinical utility when used alone. In the study by Dong et al., the optimal cut-off value for bacilli was reported as 50, with a sensitivity of 69.5% and an NPV of 82%. In contrast, we identified a higher optimal cut-off value for bacilli, while maintaining a broadly comparable diagnostic profile, suggesting potential variability related to device-specific characteristics and population differences (Dong et al., 2022).

When the overall screening performance was evaluated, WBC demonstrated slightly higher discriminative ability compared to bacillus, total bacterial count, and suspected coccus. In contrast, Wang et al. reported superior performance for bacterial parameters (AUC: 0.86) compared to WBC (AUC: 0.74) using a different analytical system, and suggested that combining leukocyte and bacterial parameters could improve diagnostic performance (Wang et al., 2023). Consistent with this observation, our findings indicate that combining WBC and bacillus parameters enhances screening performance.

Millan Lou et al., in a comparative study of two different systems, reported that bacterial counts demonstrated higher diagnostic performance than leukocytes, and that combining bacterial and leukocyte parameters improved overall accuracy (Millán-Lou et al., 2018). In line with these findings, the combination of WBC and bacillus in our study resulted in improved specificity and negative predictive value, albeit with a reduction in sensitivity, reflecting a trade-off inherent to screening strategies.

Similarly, Koçer et al. reported comparable diagnostic performance between bacterial and leukocyte parameters (AUC: 0.710 and 0.728, respectively), which is consistent with the similar performance observed in our study (Kocer et al., 2014). In contrast, Haugum et al. reported substantially higher AUC values for both bacterial and leukocyte counts (0.94 and 0.81, respectively), highlighting variability across different analytical platforms and study populations (Haugum et al., 2021).

The combined use of WBC and bacillus provided the highest overall screening efficiency in our study. Notably, subgroup analysis revealed superior performance in male patients compared to females. This difference may be explained by higher contamination rates in female urine samples due to anatomical factors, leading to reduced specificity and screening efficiency. Similar sex-related differences in screening performance have been reported in previous studies (Ilki et al., 2017; Dong et al., 2022; Kouri et al., 2024). These findings suggest that the use of gender-specific cut-off values may improve diagnostic accuracy in routine practice.

In male patients, bacillus demonstrated particularly high specificity and positive predictive value at optimized cut-off levels, indicating that this parameter may be especially useful for ruling in infection and guiding early antimicrobial therapy decisions when Gram-negative UTI is suspected. In contrast, suspected coccus showed limited diagnostic value, with relatively low predictive performance. Hua Wang et al. similarly reported that Gram-negative flags demonstrated high positive predictive value, whereas Gram-positive indicators were less reliable, further supporting the limited clinical utility of suspected coccus as an isolated parameter (Wang et al., 2023).

From a clinical perspective, an effective screening strategy for UTI should primarily aim to achieve high sensitivity and negative predictive value in order to safely exclude infection. According to European Urinalysis Guidelines, screening methods for clinically significant bacteriuria are expected to achieve sensitivity levels of 90%–95% (Kouri et al., 2024). In our study, although high negative predictive values were achieved, sensitivity levels remained below the recommended range in the overall population, with the exception of male patients, indicating that while the system is effective for ruling out infection, caution is required when using it as a standalone diagnostic tool.

The MUS-3600 analyzer generally demonstrated moderate diagnostic performance. In the study cohort, the highest overall elimination rate (77.2%) and negative predictive value (87.9%) were obtained when WBC (>39/μL) and bacilli (>103/μL) were evaluated together. These results suggest that unnecessary cultures can be reduced in certain cases. However, the relatively low sensitivity (44.4%) and positive predictive value (34.5%) observed in the classification of suspicious cultures limit the reliability of this parameter when used independently.

Subgroup analysis showed better performance in male patients with an elimination rate of 86.9% and an NPV of 93.9% compared to female patients with an elimination rate of 68.9% (NPV: 82.1%). This gender-related variability indicates that diagnostic efficiency can be improved by using specific cut-off strategies. While these findings support the use of MUS-3600 as a screening tool, none of the evaluated parameters reached the 100% NPV required for the safe exclusion of infection. In absolute terms, this corresponds to approximately 20 missed infections among 350 male patients (6%) and 61 cases among 409 female patients (15%) in our cohort. Therefore, culture decisions should be individualized, especially for high-risk populations such as immunocompromised individuals and intensive care unit patients.

In conclusion, flow-type microimaging systems such as MUS-3600 provide a valuable contribution to laboratory workflows. They are not intended to replace urine culture but rather to complement it by improving laboratory workflow efficiency. These findings need to be further validated in larger prospective studies, and it is necessary to investigate whether gender-specific cut-off values or advanced analytical approaches, including machine learning, can improve diagnostic performance.