The Application of Microfluidic Chips in Primary Urological Cancer: Recent Advances and Future Perspectives

2.1 Microfluidic Chip

Microfluidic technology can regulate fluid streams in a channel in a microenvironment. When applied to an OOC, microfluidics simulates the natural flow of body fluid, mimicking the transportation of nutrients and elimination of waste liquid. With the progress of polymer materials and microsystem processing technology, polydimethylsiloxane (PDMS) microfluidic chips in biomedicine have gradually received abundant attention [19]. The PDMS chip has decent hydrophobic permeability and biocompatibility. It can precisely control various biochemical conditions through microchannels, mimicking human microenvironments with in vitro culturing of various human tissues and organoids [20].

The analytical performance of a microfluidic chip largely depends on the quality of the microchannels created inside the chip. At present, commonly used microchannel processing technologies include laser-guided direct writing [21], photolithography [22], hot embossing [23], micro-milling [23, 24], and 3D printing [25]. Laser-guided direct writing uses laser ablation to directly generate three-dimensional microchannels in glass slides without additional stacking and bonding. Photolithography and hot embossing are indirect techniques for fabricating microchannel structures, requiring the processing of molds and then the molding of microchannels [26]. Micro-milling and 3D printing technologies can directly process micro-channels on the surface of the workpiece or create a mold and then generate the micro-channel structure. In photolithography, SU-8 negative can be used as a photocurable polymer, which is solidified on the mold to produce the chip microstructure. The microstructure is then transferred to the polymer PDMS chip using the mold [27]. Positive photoresists can help sustain their sizes and patterns as the photoresist developer solvent does not permeate the areas not exposed to UV light. With negative resistance, the solvent permeates both the UV-exposed and unexposed areas, which lead to pattern distortions. Bonding is an important link in chip fabrication. The sealing of microfluidic channels can be performed by plasma oxidation sealing [28], anodic bonding [29], ultraviolet irradiation [30], organic solvent bonding and cross-linking agent adjustment [31]. Since PDMS relies on intermolecular forces to achieve bonding, liquid leakage is prone to occur when the pressure is high. So, the sealing between PDMS and glass chips needs to be permanent. Li et al. [32] made a microfluidic chip with PDMS-glass, which bonded the poured PDMS to the glass with a flat and clean surface by a plasma method. They avoided the cumbersome processing of traditional glass chips as glass-etching is not required for plasma bonding in this method. Compared with traditional etching [33] and sputtering processes [34], this method is simple, convenient, and fast with relatively low equipment requirements. It is thus considered one of the best methods for making chips concurrently [35]. Previous research found that the shear stress induced by urine flow ranges from 0.2 to 2 dyn cm⁻² at regular flow rates, and it was reported that the bonding force can encounter the shear force in the urinary flow microenvironment [36] (Table 2).

2.2 From Microfluidic to OOC

OOC is a system with engineered or original tissues grown in in vitro microfluidic equipment. The microfluidic chips are designed with optimized biochemical conditions which is similar to human physiology, mainly to regulate cell microenvironments and major tissue-specific functions on a real-time basis [37]. It applies microfluidic chip technology in histology and is a potential substitute for future human in vitro organ models [38]. Structurally, an OOC consists of four major functional parts: an organ tissue culture chamber, a microfluidic chip, components for stimulation or drug delivery, and biosensors [39]. In addition to the advantages of tissue function and physiological structure, current OOCs could also enable the manipulation of fluids and the dynamic capture of parameter changes on the micrometer scale, further improving the simulation degree of biological models and the sensitivity to delicate experimental parameter changes.

2.3 The Special Cell Culture Style of OOC

Compared to conventional 2D cell culture, 2.5D and 3D cultures are most prevalent in OOC technology. They are both in a dynamic environment and usually have a variable fluid environment, while 2.5D culture mainly culture cells on the basement membrane [40]. Due to technical support and cost limitations, 2.5D culture is still primarily used in OOCs [41].

2D culture is a single layer of cells, and the 3D culture is generally an ellipsoid-shaped cell mass. The microfluidic model of monolayer cells based on 2D culture is simplified to evaluate one or two layers of cells in a fluid environment [42]. The 3D ellipsoid in the microfluidic is more similar to the natural tissue and can better reflect the real tumor microenvironment [43]. One specific kind of OOCs provides a double-layered membrane for two kinds of cells growing in a 2.5D environment, which allows convenient observation of the cells' connection and interaction in the pulmonary alveoli and the blood vessels [17, 44]. When the organoids with different cells were connected in parallel or series flow channels, those organoids refer to 3D structures grown from stem cells with induced organ-level functions, usually in a more static microenvironment, and are slightly different from the organ-specific cells used in most of the OOC. Overall, OOCs are powerful tools for studying functional units of human organs in microfabricated cell culture devices [45].

2.4 Micropumps of OOC

The micropump powers the flow of culture media in the microenvironment, and fluid control is mainly driven by micropumps at the cellular level [46]. The micropumps can be classified into two types: mechanical and non-mechanical. Mechanical micropumps include pneumatic micropumps, electromagnetic micropumps, centrifugal micropumps [47], and piezoelectric micropumps. Non-mechanical micropumps include gravity-driven pumps and electroosmotic-driven pumps [48]. Currently, most simple fluid control systems use injection pumps [49] and peristaltic pumps [50] for sample transport and expulsion. For instance, the pressure of urine flow is one of the mechanical characteristics of the microenvironment of human urethra. The urine flow during the emptying process can be compared to an injection pump operating under appropriate pressure. When urine is formed and pushed from the upper urinary tract to the bladder, it could be seen as a dynamic peristaltic pump operating at the appropriate pressure. These pumps are also used to power the movement of nutrient cells needed in the microenvironment to accomplish dynamic cell culture. Dynamic liquid perfusion is another distinctive feature of microfluidic culture compared to traditional 2D culture.

2.5 Biosensors in OOC

At present, the activity of cells in the chip is mainly monitored using fluorescent biomarkers, whose signals could be real-time captured for organoids and subsequent experimental operations [51]. With biosensors, the overall growth state of cells can be continuously observed. The electrical impedance sensor detects the status change of the biological element through the change of electric potential without additional labeling. An OOC developed by Osaki et al. simulated the interaction of neuromuscular junctions in 3D, and was able to record axonal growth, neuromuscular junction formation, real-time optogenetic regulation of muscle contraction, and motor neuron activity [52]. The results strongly suggest that various neural pathological behaviors associated with muscle disease can be simulated in OOCs. In addition to the detection of potential signals, Zhang et al. developed a highly integrated modular OOC platform using the characteristics of antigen-antibody binding. The platform has multiple sensors built-in, including test plates for microfluidic routing through built-in pneumatic valves, microbial reactors for organoids, physical sensing kits for measuring microenvironmental parameters, one or more electrochemical sensing units for detecting soluble biomarkers secreted by organoids, media storage and bubble traps, and fluid flows through each module. Each module in the circuit can also be replaced individually, such as the dielectric storage, bubble trap, and electrochemical sensor chip [53]. As a part of the OOC system, the existence of multifunctional, easy-to-adjust sensors makes this technology convenient to use, cost-effective, and reliable for clinical research (Table 3).

3.1 Urine Examination With Microfluidics

The clinical diagnosis and follow-up of BCa include urine exfoliative cytology, color doppler ultrasonography, CT scan, and cystoscopy [56]. Among them, urine exfoliative cytology is a standard clinical method for disease monitoring, although with relatively low efficiency [57]. Since extracellular vesicles (EVs) and other metabolites of bladder mucosa and tumors enter the urine through various pathways, they are suggested as potential clinical biomarkers in urine.

3.1.1 Urinary Exfoliated Tumor Cells (UETCs)

UETCs are derived from the cells shed off from bladder tumors and are expected to share the same genomic information with BCa cells [58]. UETCs have been a diagnostic marker in urinary cytology for over 2 decades. Nevertheless, current methods' low sensitivity and poor qualification nature (i.e., lack of assay standardization and inconsistent cut-off values) have resulted in various limitations in clinical practice [59]. Those defects significantly discouraged the application of UETCs as a biomarker for BCa diagnosis. With a high recurrent rate in BCa, traditional urinary cytology testing is not beneficial in decreasing the screening workload for detecting BCa, especially for patients with recurrent disease. To tackle these issues, novel methodologies were developed to capture UETCs. In 2013, Birkhahn et al. [60] designed a device using polycarbonate membranes with uniformly distributed perforations of 7.5 μm size to capture different types of UETCs. It employs a cell size-based capture mechanism with a dimension of 7.5 μm comparable to the size of unwanted blood cells in the assay, in which more giant tumor cells are captured and enriched on a reduced surface area. In contrast, smaller blood cells pass through a microfilter. While the specificities of microfilter-based and conventional cytology were similar (100% vs. 95.8%, N = 24), the former's sensitivity increased significantly compared to the latter (53.3% vs. 40%, N = 30). One major disadvantage, though, is that the device could be clogged by accumulated debris (both cellular and non-cellular) and consequently contaminate or even damage the captured cells. To solve this issue, in 2018, Chen et al. analyzed the sizes and distributions of UETCs in urine samples and made a microfluidic device to capture UETCs based on three common sizes (15, 30, and 60 μm) of UETCs that were commonly captured [61]. The microunits with those distinct sizes are collectively grouped in repeat patterns throughout the chamber and allow the capture of single UETCs of various sizes or clustered UETCs. With this modification, an increased cell recovery rate (80.6%) was achieved in cell loss analysis compared to that of the traditional pap smear (71.4%); a numerical representation of cell loss is provided by the RF, which is calculated as follows: RF = ([recovered cells on chip or on slide]/[total input cells] 100%). Increased UETC counts in patients with BCa (N = 79) were observed compared to the normal controls (NC) group (N = 43) with statistical significance (53.3 [10.7–1001.9] versus 0.0 [0–0.3.0] UETCs/10 mL; p < 0.0001). The performance from parallel experiments showed that the sensitivity of microfluidic immunoassay is 95% (N = 20), which is much higher compared to traditional cytology and two commercially available FDA-approved markers (BTA stat and NMP22 Bladder Chek) (35%, 65%, and 35%, respectively). The specificity is comparable to the other three assays (80% for microfluidic immunoassay, 100%, 60%, and 80% for traditional cytology, BTA stat, and NMP22 Bladder Check, N = 20). Overall, this experiment demonstrated that UETCs captured by the microfluidic device showed advantages over conventional cytology and two commercially available FDA-approved markers [61].

In 2018, Liang et al. developed a microfluidic integrated device that can separate and concentrate UETCs, as shown in Figure 3A [62]. This device is designed based on size-exclusion principles with a polycarbonate membrane with a uniform micro-pore size of 5 μm. UETCs with a diameter > 5 μm are captured for enrichment by the microfluidic device at the fluidic speed of 1 mL/min. The experiment used discarded urine samples, including 20 samples from healthy controls and 35 samples from BCa patients. According to the ROC-10 curve, when the specificity was 90%, the sensitivity of integrated filter microfluidic devices for BCa diagnosis was 77.1%. One of the device's characteristics is that it has higher integration and is suitable for rapid detection scenarios of POC. After separation and enrichment, exfoliated tumor cells (ETCs) are quantified by direct microchip Enzyme-Linked Immunosorbent Assay (ELISA). Isolated ETCs bind to anti-EGFR antibodies coupled to streptavidin-HRP. After incubation with 3,3′,5,5′-Tetramethylbenzidine, blue fluorescent signals were formed and captured using a mobile device, and the images were sent to a computer for data analysis. This method does not require centrifugation and traditional climbing analysis of the sample, and it is more convenient to transmit the color image directly to the computer, which is a trend in line with POC. However, the possibility of specific blockage issues cannot be ruled out. Although blood cells of 6–7 μm can be deformed through a small pore size, it is still more sensitive than the 7.5 μm device (77.1% > 53.3%). Therefore, integrated filtering devices have great potential in BCa screening and subsequent analysis of treatment effects.

In 2019, Khoo et al. developed a sorting device for exfoliated BCa cells (EBCCs), which can separate different epithelial-mesenchymal transition (EMT) phenotypes. Unlike physical size filtration and partition, this study relies on the fluid to sort different subtypes of cells. This program integrates microfluidic analysis based on the principle of inertial focusing, as shown in Figure 3B, for high-throughput isolation of cell size-based EBCCs subtype. It works mainly based on the centrifugal force of cells in the fluid cavity. When cells of different sizes flow in the microchannel, they will be concentrated at different positions in the transverse plane [63]. The device can quickly process clinically relevant urine sample volume (20 mL urine in 20 min). The experiment proved that the targeted EBCCs could reach 93.3% recovery rate. Since mesenchymal and epithelial cells have different shapes and deformation characteristics [64], the device can separate different subtypes of epithelial cells and mesenchymal cells into the corresponding outlet of the device. The antibodies used to isolate the biomarkers by this cell population are closely related to BCa, including anti-survivin antibodies and EGFR. Besides, EBCCs could express a variety of additional biomarkers based on their EMT status, including cytokeratin (CK) and vimentin (VIM), two protein markers with an essential role in the development and progression of Bca, and are considered to have great potential to be utilized in future clinical devices.

In addition to using the above biomarkers, glycan Sialyl Thomsen-Friedenreich Antigen (Sialyl-Tn) is also a biomarker for BCa detection. In 2020, Carvalho et al. developed a BCa detection device, UriChip, in which a microfluidic system can enrich and concentrate UETCs either in fresh urine or frozen urine [65] as shown in Figure 3C. Five rows of microcolumns with various column spacings make up the device's core structure, which filters and enriches cells in the urine. UriChip showed a 53.3% capture efficiency for Sialyl-Tn-specific UETCs. According to an in situ immunoassay, BCa patients had more sialic acid-Tn-positive cells in their urine than healthy controls (p = 0.0006). Notably, this is the first instance in which microfluidic technology has been used to combine Sialyl-Tn levels with urothelial exfoliated cells isolated from BC patients' urine. The work also showed that the contents of UETCs expressing a significant quantity of sialic acid-Tn may be extracted using frozen urine sediments. The evaluation of cryopreserved urine deposits by microfluidics is among the very first attempts. In conclusion, UriChip employs Sialyl-Tn as a biomarker to collect UETCs for tumor grading and staging in combination with the microfluidic device, offering a novel method for identifying urothelial BCa cells.

3.2 Circulating Tumor Cells (CTCs) Detection With Microfluidics

Circulating tumor cells (CTCs) are crucial for BCa detection and staging [75]. Studies using CTCs to facilitate cancer detection have grown tremendously in recent years. However, how to capture circulating tumor cells economically and efficiently is still an open question [76]. In 2020, Wang et al. created a microfluidic tool with an innovative herringbone channel structure to collect CTCs, as shown in Figure 3E. Monoclonal antibody BCMab1 against BCa was modified with biotin prior to the experiments, and human BCa cells (T24) were spiked into blood specimens from healthy volunteers to mimic circulating tumor cells in patients. The staggered herringbone pattern produces a capture efficiency of 90% (vs. a 49% capture efficiency using chips without a staggered herringbone pattern) with high specificity [77]. The capture rate of CTCs was 90% at a cell concentration of 5000 cells/mL, which improved significantly compared to the efficiency of CellSearch (28.3%) [78]. A wider groove with increased groove pitch is also suggested to improve cell capture [75]. Furthermore, since each streptavidin molecule can bind to up to four biotin molecules, they developed a multi-level signal amplifier system to maximize the signal [78]. In summary, this system could achieve increased cell capture yields, efficiency, stability and specificity.

In 2022, researchers developed an impedance-based microfluidic chip [79], as shown in Figure 3F. They generated microdroplets at the single-cell level using cancer cell lines and cells from clinical samples and introduced an integrated microfluidic electronic sensor to calculate the grade of BCa. Droplet microfluidics uses compartments that are on the same size scale as the cell, and it could combine with high-accuracy, high-sensitivity droplet analysis methods, such as fluorescence detection, mass spectrometry, electrochemical detection and surface-enhanced Raman scattering for single-cell analysis. The electrical properties of cells could facilitate the understanding of the complex physiological states of cells and help classify various types of tumor cells [80], stem cells [81], and blood cells [82] without specific labeling during the detection by electronic sensors. When it comes to BCa cell categorization, the electrical detection system can distinguish between tumor samples and normal epithelial tissue using droplets and microelectrodes [80, 83]. Individual cells may be wrapped and captured by these droplets and microelectrodes, and their impedance can be measured in a label-free and non-invasive pattern. The classification of cancer cells was conducted according to the impedance of different measurements.

3.3 BCa OOC Application

Patient-derived xenografts (PDXs) are one of the most commonly used in vivo models for drug screening. PDX models are generated by engrafting human cancer cells into immunodeficient mice. PDXs have been generated for pancreatic cancer, hepatocellular carcinoma, breast cancer, PCa, and many other cancer types with promising clinical benefits. However, the high cost, low efficiency of engraftment, and long development time increase the financial burden of patients and have limited its usage to a small scale for clinical applications. Researchers have been looking for low-cost replacement of PDX models, such as microfluidic devices that maintain BCa cells over extended periods to study patterns of drug responsiveness and possible resistance [84]. As shown in Figure 3G, by culturing PDX and clinical patient specimens, the authors confirmed that the microchambers could maintain the phenotype of primary cancer cells long enough to evaluate cancer drug responsiveness and resistance. Also, they demonstrated that microchamber cultures from specific PDX models retained patterns of drug responsiveness and resistance observed in mice by a proof-of-concept experiment. This result is encouraging and could pave the way to a reliable microfluidics-based platform for the long-term cultivation of cancer cells and screening of anti-cancer drugs.

In 2023, Xiong et al. [85] designed a high-throughput OOC for the treatment of BCa (Figure 3H). The experimenters first cultured tumor organoids in vitro, then constructed a microfluidic platform and finally combined the organoids with the microfluidic platform to establish an OOC system for drug testing. The results showed that the organoids cultured in vitro were genetically similar to the tumors of primary patients and maintained the essential characteristics of in situ tumors. The microfluidic chip has three layers (upper, middle, and lower). The upper layer is a thick PDMS membrane structure with two holes at the inlet and outlet to add and remove culture media and drugs. Then, the organoid suspension was injected into the chip with a pre-added medium for culture. A critical design feature involved streamlined guiding pillars within the microfluidic channel, which achieved three essential functions: flow rate modulation, bubble suppression, and spatial segregation of organoid spheroids into discrete chambers. Comparative validation demonstrated a strong correlation between in vitro drug sensitivity profiles (n = 8), patient-derived xenograft (PDX) model responses (n = 2), and clinical outcomes in corresponding patients (n = 2). These findings collectively indicate that the engineered organ-on-a-chip system can effectively simulate clinical drug responses and validate the differential sensitivity of patient-specific organoids to selected chemotherapeutic agents. The above results show that the constructed OOC can initially simulate clinical medication and verify the sensitivity of different patient organoids to a selected variety of chemotherapeutic drugs.

Bacillus Calmette-Guerin (BCG) is a common intravesical immunotherapy drug used in patients with BCa. BCG stimulates innate immune and inflammatory cancer responses to eliminate BCa cells [86]. In 2021, starting with immunotherapy, Kim et al. constructed a BCa OOC built upon BCG [87] capable of simulating the tumor microenvironment (TME) in vitro, as shown in Figure 3I. In this experiment, 3D printing technology was used to combine cells with gelatin methacryloyl (GelMA) hydrogel to print into cell blocks. The cell source was based on T24 and 5637 cell lines, including MRC-5, HUVEC, and THP-1 cells. Then, the printed cell blocks were packaged into a microfluidic chip. The results showed that the cells in the round cell block could remain alive under the condition of 20 μL/min microfluidic flow and 15% cell filling density in the hydrogel. The cell migration increased with the BCG dosage, while cancer cell proliferation was inhibited with the increase of BCG concentration. This study assessed the cell survival and proliferation rate, chemotaxis of monocytic THP-1 cells, and concentration of cytokines following treatment with BCG in order to determine the immunologic effects of BCG in BCa OOC (BCOC). The cell viability of BCOCs decreased in a dose-dependent manner after BCG treatment and was statistically significant in the 5637 BCOCs after BCG treatment at 1, 10, and 30 MOI (59.4 ± 1.3, p < 0.05; 52.7 ± 1.0, p < 0.05; and 20.6 1.3, p < 0.01, respectively). Based on these findings, the chip can potentially facilitate immunotherapy for BCa by determining the dosage of BCG in patients. Due to the use of cell lines and hydrogels used in common studies as substrates, this study did not simulate the specific cells of real individuals' blood and urine environment, and further studies are required to justify such simulations in vitro (Table 4).

4.1 PSA Detection With Microfluidics

PSA is an efficient PCa biomarker commonly used in PCa screening [3]. PSA exists in both free status and bound with other proteins in the blood. However, traditional PSA tests have disadvantages, such as high sample consumption and restricted analyte control in the reaction chamber. Using microfluidic devices is thus expected as an effective method to overcome these deficiencies. Currently, automated analyzers are used in centralized facilities with high throughput for PSA testing. This procedure is time-consuming and needs extensive infrastructure, including strict sample transportation regulation and professionally trained medical personnel. In 2019, Maj-Hes et al. [89] developed a microfluidic-based fingertip blood PSA rapid quantitative analysis system named Claros. Claros analyzer consists of a disposable microfluidic test cartridge (approximately the size of a credit card) and a compact analyzer (12 × 9 × 20 cm). The assay is founded on the principle of a silver amplification immunoassay. The sample and reagents travel through the 50-micron microfluidic channels during the analysis. Antigens of interest (PSA) are captured in predetermined regions (detection zones) of the channels, which apply microfluidic technology to perform POC detection directly from fingertip blood. In the experiment, the researchers compared the PSA levels in the experimental samples (100 males asymptomatic for prostate disorders) with the data from two commercially available PSA tests conducted by a reference laboratory (Abbott and Elecsys by Roche). The results showed a high correlation, indicating that Claros has the advantage of instant detection while ensuring effectiveness.

Total PSA, free PSA, and intact PSA have all demonstrated reasonably strong predictive effects in PCa [90]. However, PSA screening based on concentration alone has been linked to an increase in low-grade PCa diagnosis, which has an almost 100% 5-year survival rate [91]. Due to the lack of specificity for high-grade PCa, positive PSA screening results often result in overdiagnosis, overtreatment, and unnecessary follow-up operations [92]. Thus, there is an urge demand for PCa screening with higher specificity of high-grade PCa, which has a much worse prognosis than low-grade PCa. Increasing evidence suggests that patients with PCa have higher levels of α-2,3 sialylation and fucosylation than those with benign prostatic hyperplasia [93]. In addition, differentiating between high-grade and low-grade PCa using changed PSA glycosylation patterns may benefit patients with PCa [94]. Since glycoprotein sialylation and fucosylation are differentially regulated [95], a combination analysis of serum PSA fucosylation and 2,3-sialylation will likely increase the specific detection of high-grade PCa. In 2022, Hatano et al. [96] used microfluidic electrophoresis to measure α2,3-Sia-PSA and α-1,6-Fuc-PSA using a microfluidic immunoassay system [97]. 2,3-Sia-PSA and 1,6-Fuc-PSA concentrations were measured using a microfluidic immunoassay system based on the liquid-phase binding electrokinetic analyte transport assay principle [97, 98]. In the 2,3-Sia-PSA assay instrument, fluorescence signals from laser-induced fluorescence detection were analyzed by software designed to use internal fluorescent markers to align and identify the peaks of 2,3-Sia-PSA and 2,6-Sia-PSA (Figure 4A). The serum samples were thawed and simultaneously analyzed. The time between blood collection and storage in the freezer was approximately 3 h. In the α2,3-SiA-PSA detection system, the peaks of α2,3-SiA-PSA and α2,6-SiA-PSA were detected and analyzed based on laser-induced fluorescence. Their data suggested that the percentage of α2,3-Sia-PSA and the percentage of α1,6-Fuc-PSA were significantly higher in patients with high-grade PCa than in those with negative biopsies or with low-grade PCa (p < 0.0001). These findings suggest that a combination model of 2,3-Sia-PSA and 1,6-Fuc-PSA could effectively identify patients with high-grade PCa and distinguish them from those with low-grade PCa or from health controls. Large-scale prospective studies are required to validate these findings.

4.2 Microfluidics and PCa Exosome Detection

Exosomes, with diameters ranging from 30 to 200 nm, can be produced in multiple kinds of body fluids such as blood, saliva, and urine [99]. Exosomes' molecular contents represent the particular physiological circumstances and roles of their original cells [100]. Due to their high concentration and steady circulation status, tumor-derived exosomes are considered as viable liquid biopsy biomarkers for cancer patients. Most approaches to isolate exosomes depend on general physicochemical characteristics, such as particle size and density. The most popular technique for exosome isolation is ultracentrifugation, which requires long-time centrifugation (> 8 h), but the yield and purity are usually poor [101]. A microfluidic Raman biochip for in situ exosome extraction and analysis was created by Wang et al. in 2020 [102] as shown in Figure 4B. CD63 antibody is used for the capture of exosomes due to its high expression in exosomes and EpCAM bound Raman beads were used to detect PCa specific cells due to the significantly different expression levels of EpCAM in exosomes from PCa cells and normal cells. The procedure is designed as follows: firstly, anti-CD63 magnetic nanoparticles (Mag-CD63) that can secure exosomes in a particular position are created. Secondly, an EpCAM-Raman bead with a high density (EpCAM-Raman pearls) was created to facilitate signal detection. The specific principle is depicted in the diagram: Mag-CD63 and the detected exosomes were injected into the first and third ports, respectively, and thoroughly mingled through a rectangular structure; Phosphate buffered saline (PBS) was administered through the second port. EpCAM-Raman beads were rinsed and injected into the fourth port to blend them in the subsequent circular Raman detection area, and washed by PBS again. 20 μL of exosome samples could be analyzed per hour by Raman spectroscopy. The analysis of clinical samples verified that more exosomes were isolated from the serum of PCa patients compared to that of healthy individuals. The average number of exosomes in the serum of PCa patients was calculated to be 10.61 × 10⁸ particles/mL, approximately three times that of healthy controls. The authors conclude that microfluidic Raman devices can effectively differentiate between PCa patients and healthy controls (p < 0.0001).

4.3 Microfluidics and PCa Cell Staging

The initial development of PCa is often androgen-dependent, so androgen deprivation therapy (ADT) has always been the standard care and first-line therapy for androgen-dependent PCa. Unfortunately, most PCa patients eventually relapse into the androgen-independent growth stage, which is resistant to ADT and usually associated with poor prognosis [103]. Since cancer migration and invasion potential is highly correlated with their mechanical strengths, it was suggested that measuring the mechanical strength for sample cells might be indicative of their metastatic potential [104]. In 2019, Liu et al. conducted morphological-rheological microfluidic-based high-throughput mechanical phenotyping of androgen-sensitive and insensitive human PCa cell lines [105]. Although this morphological rheology has been studied concerning blood cells, it is yet unknown how it will be employed in epithelial carcinoma cells as shown in Figure 4C. In this study, a high-speed camera was combined with a microfluidic chip of 25 × 25 × 300 μm (width × height × length). Images were captured and analyzed under different flow speeds, and the mechanical strength of the sample cells was calculated. Thus, the androgen-sensitive epithelial tumor cells (LNCaP, DU145, and PC3) at different stages can be distinguished. The results showed that the elastic modulus of LNCaP cells, DU145 cells and PC3 cells were 1.08 kPa, 1.44–2.4 kPa, and 1.87–2.40 kPa, respectively, suggesting significantly different mechanical features of those cells. The mechanical strength of LNCaP cells was the lowest, followed by DU145 cells, and the mechanical strength of PC3 cells was the highest, supported by the measurements by Atomic Force Microscopy (AFM). Therefore, compared with ADT, this technique has a potential for early detection of PCa based on mechanical strength. Compared to the high cost associated with using an atomic force microscope, using microfluidic technology and a high-speed camera lowers the cost of mechanical characterization of cancer cells.

4.4 Detection of Prostate CTCs

Non-invasive liquid biopsies have become more prevalent for early cancer detection, monitoring and prediction of drug responses in recent years. CTCs have emerged as biomarkers to provide genetic and phenotypic information during cancer evolution from primary sites to metastatic sites [106]. Current separation methods are based on the size of different cells in the blood [107] yet the purity of separated CTCs and operation simplicity are limited. In 2017, Renier et al. designed a device that uses an eddy current microfluidic chip to isolate CTCs from the blood of PCa patients [108], as shown in Figure 4D. The device tracks the movement of cells in the microfluidic channel to screen and enrich relatively large cells in a high-speed flowing fluid (8 mL/min). In 2019, Obayashi et al. updated the microfluidic device with lower cost and better sensitivity, and the chip is composed of two different types of micropillar arrays. T. Ohnaga et al. [109] further modified the device to prevent whole blood from hindering the channels. The distance between micro-posts was increased to 200 mm near the chip inlet. The device was coated with mouse anti-human EpCAM antibody (primary) and goat anti-mouse IgG antibody (secondary). The polymer CTC-chip immobilized with surface antibody was placed in a holder that allowed a liquid sample to flow through a channel. The two apertures of the holder were then connected to a syringe pump and a sample tube. The capture efficacy was determined by comparing the number of cells remaining on the chip after sample passage to the number of cells that entered the chip inlet. To count CTCs in patients with metastatic PCa, they used a CTC-capture polymer chip. The average capture rate of PC3 cells in PBS was 94.60%, and the average capture rate in whole blood was 83.82%. The average capture rate of LNCaP cells in PBS was 82.73%, and the average capture rate in whole blood was 75.78%. In 2020, Cho et al. [110] created a lateral magnetophoretic microseparator for CTC separation to better understand the malignancy features, as shown in Figure 4E. The performance was compared between the lateral magnetic induction micro-separator (“CTC-µChip”) and the commercially available specialized method AdnaTest ProstateCancer (Qiagen) for the isolation of CTCs from the blood of PCa patients. The results showed that CTC was detected in 14/14 (100%) patients using CTC-μChip and 9/10 (90%) patients using AdnaTest. The average CTCs separated by CTC-μChip and AdnaTest were 14.8 and 0.83 CTC/mL, respectively. The contrast is even sharper in samples from patients with metastatic PCa, with 20.54 versus 2.29 CTCs/mL by CTC-μChip and AdnaTest, respectively. In conclusion, the average number of CTCs isolated by CTC-μChip was higher than that of CTCs isolated by AdnaTest.

Real-time, non-invasive methods for genomic analysis of tumors are of specific interest largely due to the highly variable disease progression and treatment response in advanced PCa. Due to their propensity to be informative through enumeration and RNA expression, CTCs have a unique potential as therapeutically effective biomarkers in liquid biopsy. Kozminsky et al. used the microfluidic chip, which was originally developed by Yoon et al., to study prostate CTCs [111]. The captured CTC extracted RNA from the parallel microfluidic chip, as shown in Figure 4F, to determine the CTC features associated with the progression and survival of advanced PCa. CTCs were isolated from the whole blood of healthy individuals (n = 8) and patients with metastatic castration-resistant PCa (n = 41). The median CTC count was 20 CTC/mL in PCa samples (3-166 CTC/mL), and that in healthy controls was 3 CTC/mL (0-14 CTC/mL), with a statistically significant difference (p = 0.0001). This study demonstrates that the CTC separation device can continue to conduct in-depth analysis of the patient's CTC in combination with single-cell techniques such as sensitivity-based RT-qPCR technology [112] (Table 5).

5.1 RCa Cell-Free DNA (cfDNA) Detection With Microfluidics

Elevated amounts and fragmentation of circulating cfDNA in plasma have been clinically detectable [115], which has been helpful for the diagnosis, prognosis, and monitoring a range of malignancies [116]. Studies have shown that the length of cfDNA fragments might have a correlation with the progress of cancer [117]. In 2018, Yamamoto et al. [118] took the fragment size of cfDNA into consideration when referring to the relationship between the overall level of cfDNA and RCa. A microfluidic-based platform was then used to measure the size of cfDNA fragments using plasma samples from RCC patients (n = 92) and healthy individuals (n = 41). The results of real-time PCR detection of these samples showed that the plasma cfDNA level of RCC patients was higher, and the fragment size of plasma cfDNA from RCC patients was shorter, when compared to that of healthy controls. However, the two groups had no statistical difference (p = 0.052). Interestingly, in patients with hepatocellular carcinoma, the size distribution of plasma cfDNA is transferred to shorter fragments as the proportion of tumor-derived DNA increases [119]. These studies demonstrate the strong potential of cfDNA fragment size to facilitate RCC diagnosis.

5.2 Renal OOC

Anti-angiogenic RCa therapy is an essential means to inhibit tumor growth in cancer treatment. However, optimization of the dosage of anti-angiogenic drugs is challenging in different patients, and the response to different drugs varies significantly in individuals [120]. Therefore, more efficient in vitro models are required for drug screening studies of patients with RCa. In 2019, Jiménez-Torres et al. used a microfluidic-based in vitro model in renal cell carcinoma anti-angiogenic drug response testing [121] in which normal endothelial cells (NEnC) and primary patient-specific tumor endothelial cells (TEnCs) were used to create patient-specific biomimetic blood arteries, as shown in Figure 5A. A LumeNext microdevice method was used to fabricate vascular OOCs in this experiment [122], which is similar to casting technology. The hydrogel is poured into a cavity filled with rod-like PDMS material, and then solidified completely. After removing the PDMS rod, a hydrogel tubular cavity is obtained. The extracted tumor and regular cell suspensions were added to the microchip for culture. In the subsequent drug response experiments, paired normal and cancer tissues from clear cell renal cell carcinoma (ccRCC) patients (n = 5) were combined with microchips to form an in vitro model of patients, and three different concentrations (10 nM, 500 nM, 1 μM) of anti-angiogenic drugs were used to treat bionic blood vessels [123]. Surprisingly, only some patients' in vitro test models responded to anti-angiogenesis therapy. This finding is consistent with the clinical observation of response heterogeneity and the development of resistance to anti-angiogenic therapy [124]. Therefore, these models can potentially evaluate personalized responses to different anti-angiogenic drugs and facilitate the clinical decisions.

In the drug testing of RCa, the development of organoids has gained great interest due to the urgent need of reliable preclinical models for drug sensitivity and dosage evaluation [125]. However, current patient-derived organoid models are still insufficient for extensive drug screening. One concern is that the organoid model must be identical in each passage, consistent in size, and able to compare the tested drugs and drug interactions previously shown in primary tumors [126]. To tackle this issue, Ozcelik et al. constructed a microfluidic system to form organoid microspheres with controllable size in alginate hydrogel [127]. This study used a simple and low-cost microfluidic device to evaluate RCa cells and mesenchymal stem cells (MSC) in drug screening. The kidney-derived human RCa cells Caki and MSC isolated from human umbilical tissue were combined with sodium alginate to construct RCa cell organoids. A five-entry microfluidic device was used to prepare organoids as shown in Figure 5B. Alginate, 0.1 M CaCl₂ solution, mineral oil with 5 wt% SPAN 80, and cell solution were infused from the five inlets of the microfluidic device at 10, 20, 20, 50 and 50 μL/min to form uniform size alginate hydrogel beads containing cells. The organoid beads were placed in the medium for 3D culture. After 21 days, the cell mass was cultured into spheres, and the organoid structure was observed by sectioning. Cisplatin was used as a chemotherapeutic agent for comparison in the drug test. This proof-of-concept experiment confirmed that microfluidics and the renal organoid model are suitable for kidney cancer-related drug testing.

Cancer metastases are responsible for 90% of cancer deaths [128]. Chemotherapy is the major treatment option for patients with metastatic cancer, yet the survival rate still remains low after the treatment [129], which leads to urgent needs for drug testing models for patients with metastatic RCa. In 2020, Wang et al. produced an OOC that can simulate the process of primary RCa cell invading/metastasizing to liver [130]. The chip was based on rat decellularized liver matrix (DLM) and GelMA scaffolds. Seven different proportions of RCa cells (Caki-1 cells, or C) and hepatocytes (HepLL cells, or H) (100% HepLL; 9:1 H/C; 7:3 H/C; 5:5 H/C; 3:7 H/C; 1:9 H/C and 100% Caki-1) were co-cultured in separate chambers, as shown in Figure 5C. The chip material was provided with oxygen permeability by PDMS, and the response to 5-fluorouracil (5-FU) on RCa cells was tested. The results showed that 5-FU-loaded PLGA-PEG nanoparticles were significantly more effective than free 5-FU in removing Caki-1 cells (p < 0.05), indicating that the model is able to predict drug responses and optimize drug dosage on metastatic cancer cells.

Shiga toxin-producing Escherichia coli (STEC) can cause infection in the human intestine, which can be treated by Ciprofloxacin (CIP). However, CIP treatment would increase the secretion of Shiga toxin 2 (Stx2) and result in damage to the human kidneys and lead to hemolytic uremic syndrome (HUS) [131]. To study and observe the phenomenon of infection in vitro, in 2021, a multi-organ chip combining the intestine and kidney was developed by Lee et al. [132]. It was a gut-kidney axis (GKA) OOC and used intestinal cells (Caco-2) and renal cells (HKC-8). The results showed that the OOC could detect STEC infection and Stx2 secretion/poisonous effect. The chip was used to verify the effectiveness of gentamicin in treating STEC infection and prove that gentamicin can suppress the harmful effect on renal cells during treatment compared to CIP (Table 6).

6.1 Limitations

The review summarizes recent studies of microfluidic and OOC technology in urinary tumors. Currently, there are still some limitations that might hinder the application of those technologies in patients with urinary system neoplasms.

Various stages of BCa necessitate distinct treatment approaches. The staging of BCa is closely correlated to circulating bladder tumor cells. The use of microfluidic technology makes it possible to collect BCa CTCs efficiently. It can optimize the study of the correlation between BCa CTCs and staging, combine the different electrical impedances of BCa cells at various stages, and sort cancer cells by combining microfluidic technology with sensors. BCa OOCs maintain the in vitro culture of primary cancer cells or tumor-like organs, and uncomplicated drug response and immune response experiments can be conducted simultaneously to demonstrate its efficacy. These OOCs are limited in their ability to simulate the tumor microenvironment based on the time and space of cancer cell culture, and there is still much opportunity for improvement. (Tables 7 and 8).

In the study of PCa, microfluidic technology is utilized primarily to evaluate biomarkers of PCa cells, including the quantification of PSA, the detection of exosomes, and the separation and enrichment of prostate CTCs. PSA is recognized as an effective tumor marker: a high sample consumption, a single analysis, and a high price limit traditional PSA detection. Microfluidic technology provides a versatile platform for laboratory research that can be optimized for POC equipment portability. In order to respond to high-level PCa research, scenarios or flexible combinations of analysis have been increased to quantify various PSAs. A second study on the staging of PCa cells demonstrated that microfluidics can be combined with high-speed cameras to stage PCa cells in various stages of epithelial tumors. In this instance, morphological rheology analysis is utilized. The molecular content of exosomes in the prostate can reveal the specific physiological environment and function of its original cells, such as EpCAM in PCa and non-cancer exosomes. EpCAM in exosomes of PCa cells and non-PCa cells differ significantly. Using microfluidic technology in conjunction with magnetic nanoparticles, a microfluidic hybrid platform for Raman detection of tumor cells was developed in this study. It assures the precision and integrity of detection without destroying the body of the chip, demonstrating the potential of microfluidic chips when combined with other technologies. CTCs have emerged as a popular biomarker that could provide genetic and phenotypic characteristics representative for primary and metastatic tumor sites. It has been compared to the FDA-approved CellSearch System for separating CTCs by microfluidic chip, but it also has some shortcomings. The current research focuses on optimizing the chip's cost, time, and detection efficiency, producing fruitful outcomes. Other studies on CTCs using microfluidic devices and RT-qPCR have also served as a source of inspiration.

Microfluidic chips for RCa differ from those for BCa and PCa, as it focuses on OOC technology to create an in vitro model to simulate the RCa tumor microenvironment instead of enrichment and evaluation of specific biomarkers. Cell-free DNA (cfDNA) has aided in the diagnosis, prognosis, and progression monitoring of RCa, as studies have linked the length of cfDNA fragments to carcinogenesis. Researchers have constructed various forms of OOCs or organoid microspheres for anti-angiogenesis therapy and drug screening using microfluidic technology, as well as created a multi-organ kidney device to reproduce the spread of kidney cancer to the liver and the role of toxins between intestinal cells and kidney cells. These studies demonstrate the requirement of RCa chips and renal OOCs in developing multi-organ chips.

6.2 Future Perspectives