This is an open access article under the terms of a license that permits non-commercial use, provided the original work is properly cited. © 2023 The Authors. Société Internationale d'Urologie Journal, published by the Société Internationale d'Urologie, Canada. SIUJ.ORG SIUJ • Volume 4, Number 4 • July 2023 Key Words Competing Interests Article Information Urinary bladder neoplasms, diagnosis, spectroscopy, Raman spectroscopy, infrared spectroscopy None declared. Received on January 15, 2023 Accepted on March 17, 2023 This article has been peer reviewed. Soc Int Urol J. 2023;4(4):321–334 DOI: 10.48083/NCZW3015 321 REVIEW Point-of-Care Diagnosis of Bladder Cancer With Vibrational Spectroscopy: A Systematic Review Arthur Yim,1,5 Matthew Alberto,1 Varun Sharma,2,3 Bayden Wood,4 Jaishankar Raman3 Lih-Ming Wong,1 Joseph Ischia,1 Damien Bolton1 1Department of Urology, Austin Health, Australia 2Department of Surgery, The University of Melbourne, Australia 3Department of Cardiac Surgery, Austin Health, Australia 4Centre for Biospectroscopy, School of Chemistry, Monash University, Australia 5Young Urology Researchers Organisation (YURO), Melbourne, Australia Abstract Introduction Vibrational spectroscopy (VS) is a new and rapidly evolving technology in cancer diagnostics. Originating from analytical chemistry, VS evaluates vibrations of nuclei to produce a unique “biological fingerprint.” While multiple studies have been published on this technology and physician awareness has increased, no systematic review has evaluated the role of VS in bladder cancer (BCa) tissue diagnosis. Methods To conduct this systematic review, we searched the MEDLINE, Embase, and Cochrane databases for studies that used Raman spectroscopy (RS), surface-enhanced RS (SERS), infrared spectroscopy (IR) or near-infrared spectroscopy (NIRS) to analyze human BCa specimens. Studies using animal tissue or liquid biopsies were excluded. We synthesized the evidence by comparing modalities, study design, data analysis techniques, and diagnostic accuracy. The quality of evidence was evaluated by the QUADAS-2 tool. Results Out of 362 results, 20 studies met our inclusion criteria. There has been growing interest in VS use in BCa, with 50% of the studies published in the past 5 years. RS was the most commonly used modality (65%), followed by IR (20%) and SERS (10%). Only one study compared RS to IR (5%). The mean sample size was 44 patients (range, 6–214). To date, there have been only 2 in vivo studies, with the remaining ex vivo studies performed with large variation in tissue preparation, data analysis, and reporting. Advancements in fiber optic probes and machine-learning data analysis techniques, and increased computational power have improved diagnostic accuracy up to 98% sensitivity and 100% specificity. Conclusions VS shows high potential for BCa diagnosis, but there is a need for uniform reporting methods and studies with adequate sample sizes to validate the models. RS has shown promising results, with ongoing improvements in fiber optic probes allowing its integration into conventional cystoscopes. While no single VS modality has proven to be perfect, a multimodal approach is likely required to establish its value in clinical practice. Introduction Challenges in bladder cancer diagnosis Bladder cancer (BCa) is among the 10 most common cancer types worldwide, with approximately 550 000 new cases annually[1]. The recurrence and progression rates vary greatly, based on factors such as tumour grading, size, depth of invasion, and presence of carcinoma in situ (CIS). At 5 years, the recurrence rates range from 31% to 78% and the progression rates range from 1% to 45%[2]. The stage of cancer is the most important prognostic factor, highlighting http://SIUJ.org https://orcid.org/0000-0003-2045-1231 https://orcid.org/0000-0003-3722-7754 https://orcid.org/0000-0002-5008-4113 https://orcid.org/0000-0003-3581-447X https://orcid.org/0000-0002-7691-4779 https://orcid.org/0000-0003-0490-7876 https://orcid.org/0000-0001-7177-3631 https://orcid.org/0000-0002-5145-6783 mailto:damienmbolton%40gmail.com?subject=SIUJ the importance of techniques for accurately and efficiently diagnosing BCa stage in controlling disease progression. The current gold standard diagnostic method is white-light cystoscopy, followed by biopsies or transure- thral resection of bladder tumour (TURBT) for histo- pathological examination. While white-light cystoscopy is reliable for papillary tumours, it has limitations in detecting flat carcinomas such as CIS, dysplasia, and multifocal lesions. Integrated findings from 2 fluores- cence cystoscopy registration studies revealed that only 38% of CIS lesions[3] and 71% of CIS cases were detected using white-light cystoscopy[3,4]. Although newer opti- cal techniques such as fluorescence and narrow-band imaging can improve tumour visualization, they do not contribute to histopathologic diagnosis[5]. Thus, repeated biopsies are often performed, a procedure that not only is costly but also does not provide real-time point-of-care diagnosis. Delays in diagnosis and treat- ment may lead to increased morbidity and mortality, particularly for high-risk invasive BCa. A more rapid and cost-effective diagnostic method would potentially enhance management of BCa patients. Vibrational spectroscopy Spectroscopy has attracted attention in cancer diagnosis in recent years. Originating from analytical chemistry, vibrational spectroscopy (VS) is a powerful technique that measures the vibrational energy of molecules[6]. The 3 most common techniques used in cancer detection are infrared (IR), Raman (RS), and near-infrared (NIR) spectroscopy. The key characteristics, advantages, and limitations of each technique are summarized in Table 1[7]. Abbreviations ANN artificial neural network AUC area under the curve BCa bladder cancer CIS carcinoma in situ FFPE formalin-fixed paraffin-embedded FT-IR Fourier transform infrared IR infrared NIRS near-infrared spectroscopy PCA principal component analysis QCL quantum-cascade laser QUADAS-2 Quality Assessment of Diagnostic Accuracy Studies RS Raman spectroscopy SERS surface-enhanced Raman scattering TURBT transurethral resection of bladder tumour VS vibrational spectroscopy such as protein/lipid ratio, tumour characteristics, and immune cell interactions, encompassing a multi- marker approach to cancer diagnostics[13]. In recent years, proof-of-concept studies in breast, colon, skin, and bladder cancers have demonstrated that VS can be employed as a label-free, non-destructive, and non-inva- sive approach to specimen analysis, facilitating the iden- tification of specific “spectral biomarkers”[14]. Objective of this review Despite the increasing literature and public awareness into the role of VS in BCa tissue diagnosis, no systematic review has covered the topic. This review aims to • Provide a historical overview of the development of VS in BCa diagnosis. • Compare the 3 most common VS techniques—IR, RS, and NIRS. • Assess the feasibility and diagnostic accuracy of studies. • Identify future areas of research based on the current literature. TABLE 1. Overview and comparison of the three common vibrational spectroscopy methods used in tissue analysis Mid-IR NIR Raman Basis Method of detection Mid-infrared light absorption using polychromic light source Near-infrared light absorption using polychromic light source Inelastic light scattering using monochromic laser excitation Wavenumber 400–4000 cm-1 4000–10 000 cm-1 50–4000 cm-1 Wavelength 2500–25 000 nm 1000–2500 nm 2500–20 0000 nm Sampling Methods Sample interface Historically complicated Straightforward point-and-shoot, versatile Straightforward point-and-shoot Sample preparation Complex Minimal Minimal Penetration of glass / quartz / plastic No Yes Yes Probes Probes fragile Fiber optic probe compatible Fiber optic probe compatible Sample Size Area 1–2 mm Up to several cm 0.3–1 mm Depth < 15 µm < 1 mm to several mm Surface spectra from bulk material possible Sensitivity Water High Medium Low Coloured samples No issue No issue Susceptible to fluorescence Specificity Chemical High Medium High Observed Bands Fundamentals—narrow Combinations and overtones— overlapping Fundamentals—narrow IR: Infrared; NIR: Near-infrared IR absorption spectroscopy relies on the absorption of mid-IR radiation by the sample, where molecules absorb specific frequencies of light based on their unique structure. This allows for identification and quanti- fication of the molecular compound in a sample. The exact frequency required to excite a molecular vibra- tion depends on the mass of the atoms involved in the vibration and the type of chemical bonds between these atoms, which can be influenced by a molecule’s structure and chemical microenvironment[8]. RS is a complementary method based on inelastic light scattering. In this method, the sample is illumi- nated with monochromatic laser light, and the inter- actions between molecules and photons leads to the scattering of light. The energy of the scattered light reflects the molecular composition of the sample[8]. RS offers several advantages over IR spectroscopy, includ- ing less interference from water and glass, less sample preparation, and higher spatial resolution. However, RS is inherently weaker and requires longer spectral scan- ning times to achieve an adequate signal-to-noise ratio. Another significant limitation of RS is the presence of strong fluorescent signals, particularly in the analysis of organic tissues, dramatically reducing its specificity and hampering its clinical translation. Surface-enhanced Raman scattering (SERS) spec- troscopy is the newest technique that aims to overcome the limitations of conventional RS. SERS uses plasmonic substrates, such as silver or gold colloids, to amplify the Raman signal of molecules adsorbed onto the metal surface[9]. This technique holds great potential in identi- fying BCa in liquid biopsies such as urine or serum, but its use in tissue is still emerging[10]. Compared with its counterparts, mid-IR and RS, near-infrared (NIR) spectroscopy has received less historical research attention. Contrary to the mid-IR region, which relies on distinct fundamental absorption bands, the NIR region contains overlapping overtone and combination bands that have lower intensity and reduced specificity[11]. However, recent advancements in quantum mechanical calculations and computational power have greatly expanded the use of NIR in modern analytical applications[12]. NIR offers advantages such as easier sample handling, low cost, greater sample pene- tration, and rapid acquisition times. All 3 techniques—IR, RS, and NIRS—can analyze biological tissues, which comprise the superposition of biochemical components such as DNA, proteins, lipids, and carbohydrates. VS can capture the unique “biological fingerprint” of the entire sample under analysis, rather than focusing on single elements like cell morphology in histopathology or tumour DNA in assays. VS has the potential to evaluate the entire pheno- typic response of the host, including tissue changes Methods This review was performed in accordance with the PRISMA 2020 statement[15]. It was registered with the International Prospective Register of Systematic Reviews (PROSPERO #CRD42022349369), where the protocol and search strategy are available. Eligibility criteria A summary of eligibility criteria for this review, fol low i ng t he PIC O f r a me work (Popu l at ion, Inter vention, Comparison, Outcome) is detailed in  Table 2. Studies of humans with  either ex vivo  or in vivo vibrational spectral analysis of bladder tissue for the detection of cancer were included. There were no demographic restrictions. Tissue samples required analysis by VS in a laboratory or operating theatre for inclusion. Types of VS considered included RS, NIRS, and IR spectroscopy. Histopathology was required as the reference standard. Publications reporting any diagnostic capability of VS were included. There were no restrictions on language or publication date. Studies involving animal tissue, pooled cells, tissue markers, or liquid biopsies were excluded. Only peer-reviewed 323322 SIUJ.ORG SIUJ • Volume 4, Number 4 • July 2023SIUJ • Volume 4, Number 4 • July 2023 SIUJ.ORG REVIEW Point-of-Care Diagnosis of Bladder Cancer With Vibrational Spectroscopy: A Systematic Review http://SIUJ.org http://SIUJ.org articles were considered, with review articles, opinion papers, and commentaries excluded. Search strategy An online electronic database search was undertaken using the platforms of MEDLINE, Embase, and Cochrane Library. The search encompassed the entire database content. The initial search employed broad MeSH terms including “Urinary Bladder Neoplasms/” and (Spectrum analysis, Raman OR spectroscopy, near- infrared/ OR Spectrophotometry, Infrared/)while also extracting key terminology/key words from reviews and a sample of potentially relevant primary data studies. A gold test set of relevant studies was used to ensure the search terms retrieved all of the gold test set. The results of the literature search were downloaded into EndNote X9 software (Clarivate Analytics, London, UK) and exact article duplicates were removed using the duplicate tool in that software program. Subsequently, a reference review of identified articles and reviews was conducted to identif y any additional relevant articles. Grey literature was searched via guidelines from the European Association of Urology (EAU), American Urological Association (AUA), and National Institute for Health and Care Excellence (NICE) and ongoing clinical trials through ClinicalTrials.gov, The ISRCTN registry, and the World Health Organization International Clinical Trials Registry Platform (ICTRP) portal. The authors of trials were contacted for preliminary or unpublished results for potential inclusion in the review. Full search strategy and results are provided in Online Appendix 1. Selection process Following completion of the search, all identified citations were uploaded into Covidence systematic review software (Veritas Health Innovation, Melbourne, Australia) and duplicates were removed. The screening Results A total of 363 articles were identified through liter- ature search, of which 263 were excluded on screening. Of 29 full-text articles assessed for eligibility, 20 were included in this review (Figure 1). Characteristics of the included studies Table 3 provides a summary of the characteristics of the included studies, while Table 4 contains a summary of all data collected. There has been a growing interest in spectroscopy and BCa since the publication of the first study in 2004, with 10 of 20 studies published in the past 5 years. The most commonly used modality was RS FIGURE 1. PRISMA 2020 diagram of study selection Records removed before screening: Duplicate records removed (n = 71) Records screened (n = 292) Records excluded (n = 263) Reports not retrieved (n = 1)  Reports sought for retrieval (n = 29) Reports excluded: Wrong study design (n = 3) Wrong tumour type (n = 3) Wrong intervention (n = 1) Conference poster (n = 1) Reports assessed for eligibility (n = 28) Studies included in review (n = 20) S cr ee ni ng Identi�cation of studies via databases and registers Id en ti �c at io n In cl ud ed Records identi�ed from: Databases (n = 362) Medline (n = 160) Embase (n = 200) Cochrane (n = 2) Registers (n = 1) TABLE 2. Criteria for studies included in this systematic review Inclusion criteria Population Human bladder cancer tissue Investigation Vibrational spectroscopy modalities: 1. Raman (RS) 2. Fourier transform infrared (FT-IR) 3. Near-infrared (NIRS) Control Histopathology Outcomes Quantitative: Diagnostic accuracy, sample size, scan time, and excitation laser wavelength Qualitative: Study design and limitations, tissue preparation, algorithm for analysis, and the pathological groups compared Setting Intraoperative, bedside, or laboratory (65%), followed by IR (20%), and SERS (10%). No studies using NIRS were found. Two comparator studies were included. One compared FT-IR and RS on the same bladder specimens[17], while another compared a novel superficial RS fiber optic probe with a non-superficial probe[18]. Sample sizes of the studies were low, with a mean of 44 patients (range, 6–214). Only 2 studies were conducted in vivo[18,19], while the remaining studies were ex vivo (n = 18). There was considerable variation in tissue prepa- ration methods, including snap-freezing bladder speci- mens post-TURBT and subsequently thawing (40%), for inclusion was conducted in 2 phases. The first phase involved screening titles and abstracts from the initial search results. The second phase involved reviewing full-text articles based on the previously stated inclusion criteria. Both phases of screening were conducted by 2 independent reviewers (A.Y. and M.A.). In cases of unresolved disagreements, a third senior reviewer (D.B.) acted as an adjudicator. The same approach was used to screen all grey literature sources. Data collection process Two reviewers (A.Y. and M.A.) independently conducted data extraction onto a predefined extraction sheet. The extracted data were cross-checked independently. The primary outcome measures extracted for assessing the effectiveness of a diagnostic modality included quantitative measures of accuracy such as sensitivity, specif icity, overall accuracy, and area under the curve (AUC) values. Secondary outcome measures encompassed bot h qua nt itat ive a nd qua litat ive data covering study design and limitations, tissue preparation, scan time, excitation laser wavelength, data analysis technique, and comparison of pathological groups. If multiple data analysis techniques were evaluated within a study, the data described are based on the most effective technique used. When data were presented for both a training set and test/cross- validation set, the data from the test set are presented, as it reflects the performance of the test in clinical practice most closely. Study risk of bias assessment Two reviewers independently assessed each eligible study using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) tool[16]. Any areas of conflict between the 2 reviewers were resolved through arbitration involving a third reviewer (D.B.), if necessary. 325324 SIUJ.ORG SIUJ • Volume 4, Number 4 • July 2023SIUJ • Volume 4, Number 4 • July 2023 SIUJ.ORG REVIEW Point-of-Care Diagnosis of Bladder Cancer With Vibrational Spectroscopy: A Systematic Review http://SIUJ.org http://SIUJ.org immediate analysis of fresh specimens post-TURBT (30%), or placing specimens into formalin that was vari- ably reversed before spectral scanning (30%). Regarding data analysis techniques, principal compo- nent fed linear discriminant analysis (PCA-LDA) was the most commonly used technique (45%). Other anal- ysis methods included support vector machines, partial least squares linear discriminant analysis, cluster-aver- aged spectra, ordinary least squares regression, princi- pal component analysis, and artificial neural networks. The quantitative measures of accuracy varied greatly across studies. While not all studies reported sensitiv- ity and specificity, some chose to report overall accuracy and AUC. Additionally, 25% of the studies provided only descriptive analysis of tissue constituents such as proteins, lipids, DNA, collagen, and cholesterol. Comparing the performance of spectroscopic techniques Quantitative measures of diagnostic accuracy, such as sensitivity and specificity, were reported in 12 of 14 studies that used RS. The remaining 2 studies reported descriptive analysis of tissue constituents instead of accuracy[17,20]. Diagnostic endpoints varied significantly across studies, depending on the histopathological categories chosen for analysis. For example, 90% of the studies compared benign to malignant tissues, while 30% compared low-grade with high-grade urothelial cancer. This variability made direct comparisons between studies difficult. Overall, the sensitivity and specificity for detecting malignancy ranged from 71% to 97% and from 72% to 100%, respectively. In contrast, only 1 of 5 FT-IR studies reported on accuracy. Hughes et al. (2013) used support vector machines to achieve a class accuracy of 98% to 99% when distinguishing conventional urothelial cancer from rare subvariants. They did not compare with benign tissues[21]. Pezzei et al. (2013) used FT-IR micro- scopic imaging with tissue microarray technology to correlate with stained histological BCa tissue sections, opening up new possibilities for spectroscopic analyses and exploration of the molecular changes associated with histopathological morphology[22]. The remaining 3 FT-IR studies reported only on concentrations of tissue constituents. Two studies with markedly different study designs used SERS. The first study, conducted by Jin et al. (2019), compared luminal and basal-like subtypes of BCa and reported an overall accuracy of 94%[23]. That study used 50 snap-frozen specimens without any benign controls. In a subsequent study, Zacharovas et al. (2022) applied SERS to freshly excised bladder tissue and extracellular fluid. Their 3-group algorithm achieved a sensitivity of FIGURE 2. Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) analysis of the included studies Study Risk of bias Applicability concerns Patient selection Index test Reference standard Flow and timing Patient selection Index test Reference standard Crow et al., 2004 Low Low Low Low Low Low Low Crow et al., 2005 High High Low High Low Low Low de Jong et al., 2006 High Low Low Unclear High High Low Stone et al., 2007 High Low Low Unclear Low High Low Draga et al., 2010 Low High High High Low Low Low Ahmed et al., 2010 Unclear High Low Unclear High High Low Barman et al., 2012 Low Low Low Low Low Low Low AI-Musletet al., 2012 Unclear Low Low Unclear High High Low Pezzeiet al., 2013 High High Low High High High Low Hugheset al., 2013 High High Low Unclear High High Low Chenet al., 2018 Low Low Low Low Low Low Low Jinet al., 2019 High High Unclear Unclear High High Unclear Pavlovet al., 2019 Low Low Unclear Low Low Low Unclear Yousifet al., 2020 High High Low Low Low High Low Placzeket al., 2020 Low Low High High Low Low Low Corderoet al., 2020 Low Low High High Low Low Low Morselliet al., 2021 Low Low Low High Low High Low Zacharovas et al. (2022)[10] Low High Low High Low High Low Stomp-Agenantet al., 2022 High High Low High High Low Low Taieb et al., 2022 High High Low High High High Low TABLE 3. Characteristics of studies included in the systematic review Characteristic Studies n (%) Period of publication 2004–2009 5 (25) 2010–2016 7 (35) 2017–2022 10 (50) Spectroscopy modality Raman 14 (65) FT-IR 5 (20) SERS 2 (10) Raman & FT-IR 1 (5) NIRS 0 (0) Sample size < 20 6 (30) 20–50 10 (50) 50 4 (20) Tissue preparation In vivo 2 (10) Ex vivo: frozen 8 (40) Ex vivo: fresh 6 (30) Ex vivo: formalin 6 (30) Data analysis technique PCA-LDA 9 (45) PCA-SVM 2 (10) PLS-LDA 2 (10) PCA 2 (10) OLS 1 (5) PCA-ANN 1 (5) Constituents only 5 (25) Histopathological categories compared Benign vs. cancer 18 (90) Grade characterization 6 (30) Stage characterization 3 (15) Subtype characterization 2 (10) ANN: artificial neural networks; CAS: cluster-averaged spectra; FT-IR: Fourier transform infrared; LDA: linear discriminant analysis; NIRS: near-infrared spectroscopy; OLS: ordinary least squares regression; PCA: principal component analysis; PLS: partial least squares; SERS: surface- enhanced Raman spectroscopy; SVM: support vector machines. Flow and timing Reference standard Index test Patient selection Q U A D A S- 2 D om ai n 0% 20% 40% 60% 80% 100% Proportion of studies with low, high or unclean RISK of BIAS 0% 20% 40% 60% 80% 100% Proportion of studies with low, high or unclean CONCERNS regarding APPLICABILITY Low High Unclear 327326 SIUJ.ORG SIUJ • Volume 4, Number 4 • July 2023SIUJ • Volume 4, Number 4 • July 2023 SIUJ.ORG REVIEW Point-of-Care Diagnosis of Bladder Cancer With Vibrational Spectroscopy: A Systematic Review http://SIUJ.org http://SIUJ.org TABLE 4. Summary of all data collected for the included studies Study/ Year Country Patients/ Specimens Malignant/ Control Tissue preparation Spectroscopy modality/ Mean scan time Laser wavelength Data analysis technique Histopathological categories compared Diagnostic endpoint Sensitivity / Specificity Accuracy / Other Crow et al. (2004)[24] United Kingdom 72/75 3/22 Ex vivo—Snap-frozen post-TURBT Raman / 10 sec 830 nm PCA-LDA • (i) normal, (ii) cystitis, (iii) malignant • (i) low grade, (ii) high grade • (i) pTa, (ii) pT1, (iii) pT2 91% / 98% 93% / 98% 96% / 96% n/a Crow et al. (2005)[40] United Kingdom 24/29 10/19 Ex vivo—Snap-frozen post-TURBT Raman, fibreoptic probe / 5–30 sec 785 nm PCA-LDA • (i) benign, (ii) malignant 79% / 89% Accuracy 84% de Jong et al. (2006)[41] Netherlands 15/15 9/6 Ex vivo –Snap-frozen post-TURBT Raman map / 20 sec 845 nm PCA-LDA CAS • (i) benign, (ii) malignant 94% / 92% Accuracy 98% Stone et al. (2007)[20] United Kingdom 24/73 41/32 Ex vivo—Snap-frozen post-TURBT Raman / 20 sec 830 nm OLS • (i) benign, (ii) malignant n/a Constituents: actin, collagen, choline, triolein, DNA, cholesterol Draga et al. (2010)[19] The Netherlands 38/63 23/29 In vivo—Live tissue prior to TURBT Raman and PDD / 1–5 sec 785 nm PCA-LDA • (i) normal, (ii) cystitis, (iii) malignant • (i) normal, (ii) cancer • (i) normal, (ii) pTa, (iii) pT1 + pT2 71% / 87% 85% / 79% 58% / 76% n/a Ahmed et al. (2010)[17] Sudan 7/14 4/0 Ex vivo—formalin, dried, grinded, KBr additive Raman and FT-IR / Scan time n/a 1064 nm Constituents only • (i) benign, (ii) malignant n/a Constituents: proteins, lipids, nucleic acids Barman et al. (2012)[25] United Kingdom 14/28 14/14 Ex vivo—fresh post-TURBT Raman,confocal probe / 5 sec 785 nm PCA-LDA • (i) benign, (ii) malignant 86% / 100% Accuracy 93% AUC 0.91 Al-Muslet et al. (2012)[42] Sudan 11/22 11/11 Ex vivo—formalin, dried, grinded, KBr additive FT-IR / Scan time n/a n/a Constituents only • (i) benign, (ii) malignant n/a Constituents: proteins, lipids, nucleic acids Pezzei et al. (2013)[22] Austria 214 214/0 Ex vivo—formalin, dried FT-IR micro-spectroscopy / Scan time n/a n/a PCA • (i) benign, (ii) malignant n/a n/a Hughes et al. (2013)[21] United Kingdom 6/6 6/0 Ex vivo—reversal of FFPE tissue FT-IR micro-spectroscopy / Scan time n/a n/a PCA-SVM • Rare subvariants only: (i) conventional urothelial cancer, (ii) micro-papillary, (iii) stroma, (iv) lymphocyte-rich, (v) clear cell, (vi) lipoid n/a Accuracy 98% Chen et al. (2018)[26] China 10/32 21/11 Ex vivo—snap-frozen in liquid nitrogen Raman, fiber optic probe / 1 sec 785 nm PCA-ANN • (i) normal, (ii) low grade, (iii) high grade 90% / 98% LG 98% / 96% HG Accuracy 93% Jin et al. (2019)[23] China 50 50/0 Ex vivo—snap-frozen SERS / 10 sec 633 nm PCA-LDA Two subtypes: • (i) luminal, (ii) basal-like n/a Accuracy 94% AUC 97 Pavlov et al. (2016)[43] Russia 22 22/13 Ex vivo—fresh post-TURBT Raman / Scan time n/a 785 nm PCA-LDA • (i) benign, (ii) malignant 97% / 96% n/a Yousif et al. (2020)[44] Iraq 46/46 23/23 Ex vivo—formalin ATR-FT-IR / Scan time n/a n/a Constituents only • (i) benign, (ii) malignant n/a Constituents: proteins, lipids, collagen Placzek et al. (2020)[27] Denmark 44/119 53/66 Ex vivo—fresh or snap-frozen Raman and OCT / Scan time n/a 785 nm PLS-LDA • (i) benign, (ii) malignant • (i) low grade, (ii) high grade 95% / 88% 81% / 61% n/a Cordero et al. (2020)[35] Denmark 28/67 37/11 Ex vivo—fresh or snap-frozen Raman, fiber optic mapping / 3 sec 785 nm PLS-LDA • (i) benign, (ii) malignant • (i) low grade, (ii) high grade 92% / 93% 85% / 83% Accuracy 92% Accuracy 84% Morselli et al. (2021)[28] Italy 114/169 40/129 Ex vivo—fresh post-TURBT Raman, fiber optic & fluorescence & reflectance / Scan time n/a 785 nm PCA-LDA • (i) benign, (ii) malignant • (i) low grade, (ii) high grade • (i) pTa, (ii) pT1 • (i) pTa, (ii) pT2 • (i) pT1, (ii) pT2 77% / 72% 73% / 65% 65% / 71% 81% / 81% 75% / 76% Accuracy 77% Zacharovas et al. (2022)[10] Lithuania 30/58 25/28 Ex vivo—fresh post-TURBT SERS /5 min 1064 nm PCA • (i) normal, (ii) cystitis, (iii) malignant 85% / 97% n/a Stomp-Agenant et al. (2022)[18] The Netherlands 75/117 51/66 In vivo—Live tissue prior to TURBT Raman fiber optic, superficial vs normal probe / 0.5 sec 785 nm PCA-LDA • (i) benign, (ii) malignant • (i) normal, (ii) low grade, (iii) high grade 90% / 87% superficial probe 80% / 85% normal probe AUC 0.95 superficial probe AUC 0.80 normal probe Taieb et al. (2022)[45] Israel Unknown Unknown Ex vivo—reversal of FFPE tissue Raman mapping / 60 sec 561 nm PCA-SVM • (i) benign, (ii) malignant 84% / 88% n/a ANN: artificial neural networks; ATR: attenuated total reflection; AUC: area under the curve; CAS: cluster-averaged spectra; FFPE: formalin-fixed paraffin-embedded; FT-IR: Fourier transform infrared; HG: high grade; LDA: linear discriminant analysis; LG: low grade; n/a: not available; NIRS: near-infrared spectroscopy; OLS: ordinary least squares regression; PCA: principal component analysis; PDD: photodynamic diagnosis; PLS: partial least squares; SERS: surface-enhanced Raman spectroscopy; SVM: support vector machines; TURBT: transurethral resection of bladder tumour. 329328 SIUJ.ORG SIUJ • Volume 4, Number 4 • July 2023SIUJ • Volume 4, Number 4 • July 2023 SIUJ.ORG REVIEW Point-of-Care Diagnosis of Bladder Cancer With Vibrational Spectroscopy: A Systematic Review http://SIUJ.org http://SIUJ.org multimodal approach improved accuracy up to 90%[28], the additional time and resource expenditure is signifi- cant, limiting this approach in a clinical setting. Comparison of VS techniques In the context of BCa diagnosis, RS has received the most attention in recent years. A systematic review of 9 original studies conducted between 2004 and 2015 demonstrated an impressive pooled diagnostic sensitivity of 94% and specificity of 92%[29]. However, the studies included in this meta-analysis were highly heterogeneous in terms of sample type, instrument used, excitation wavelength, and algorithm for analysis. Some studies used snap-frozen or formalin-fixed paraffin- embedded (FFPE) tissue, while others used cell lines, peripheral blood, or urine. There has been great interest in using RS to analyze urine samples for molecular signatures associated with BCa to develop a truly non-invasive screening test. Huttanus et al. (2020) developed an RS-based chemom- etric urinalysis (Rametrix) as a direct method for screen- ing urine samples. Using a model built with 22 principal components, BCa was detected with 82.4% sensitivity and 79.5% specificity[30]. The reduced accuracy of this label-free method could be attributed to the diversity of urine composition, concentration, and pH[31]. With the surface enhancement provided by SERS, greater diag- nostic accuracy could be achieved, as demonstrated a recent study by Hu et al. (2021) with 100% sensitivity and 98.85% specificity[32]. FT-IR analysis of bladder washings has also been proposed as a sensitive, rapid, non-destructive, and operator-independent analytical diagnostic method for BCa compared with traditional urine cytology. In a study by Gok et al. (2016), bladder washings were analyzed from 136 patients, demonstrating a sensitivity of 100% but modest specificity of 73.5%. Interestingly, traditional urine cytology had a sensitivity of only 45% on the same specimens in this study[33]. The combination of FT-IR with microscopy has also led to the development of IR imaging. Studies have demonstrated accuracies > 90% compared with immu- nohistochemical (IHC) diagnostics by pathologists[34]. However, clinical translation of this powerful integrated technique has been hindered by long measuring times and complex FT-IR setup requiring liquid nitrogen cool- ing. Quantum-cascade laser (QCL)-based microscopes have shown promise in overcoming these limitations, enabling IR imaging to be performed within minutes. Kuepper et al. (2018) demonstrated that QCL-based IR imaging could identify colorectal cancer in the same time frame as a frozen thin section diagnosis by pathol- ogists, boasting a sensitivity of 96% and specificity of 100%[34]. Limitations of evidence and review process Despite the high levels of diagnostic accuracy achieved with VS recently, this review has identified several limitations that indicate the need for further work before the clinical use of VS as a minimally invasive tool for cancer investigation. The inclusion criteria of this review did not impose limitations on sample size to provide a broader overview of all available literature on VS. As a result, one-third of studies included had fewer than 20 participants, offering a low level of evidence with significant heterogeneity in study design. Furthermore, variation in tissue sample preparation techniques, pathological grouping, and data analysis make direct comparison between studies difficult. This prevents any meaningful pooling of results through meta-analysis to obtain statistical estimates of overall diagnostic accuracies. The reporting methods of the included studies were inconsistent and often incomplete. While many stud- ies often reported sensitivity and specificity, they often omitted reporting accuracy and AUC, or vice versa. Only 1 study reported all 4 of measures of accuracy[25]. The use of the term “optimal” sensitivity in some stud- ies raises concerns about reporting bias, as the authors may have been selecting the best results for reporting. Concerningly, none of the studies provided complete data for all key areas of diagnostic accuracy: true and false positivity and negativity, sensitivity, specificity, and positive and negative predictive values. It is paramount to publish larger studies that comprehensively report these values to further evaluate spectroscopy. Implications on clinical practice and challenges for future research While many of the studies included in this review analyzed ex vivo bladder specimens, the true potential of non-invasive VS lies in its application in a real-time in vivo setting. This will undoubtedly depend upon the development and optimization of fiber optic probes that can be introduced via the urologist’s everyday cystoscope. This trend is already evident in the studies included in this review, with 4 of the 5 studies published in the past 5 years using a fiber optic probe[18,26,28,35]. Another major challenge is the presence of f luids such as urine or glycine, which can interfere with the spectroscopic signal and reduce specificity. More in vivo research is needed to evaluate the feasibility of VS in the operating theatre. The results of a phase 1 trial by Hermann et al. are awaited (ClinicalTrials.gov identifier: NCT05124106), as the study utilizes fiber optic probes to take RS measurements inside the bladder of 30 patients. It is noteworthy that no studies using NIRS to analyze BCa were included in this review, despite the success- ful use of this technique in evaluating prostate cancer, breast cancer, and cardiac fibrosis specimens[36–38]. 85% and specificity of 97% in distinguishing malignancy from cystitis and normal tissue[10]. Quality assessment and risk of bias All articles were evaluated for risk of bias and concerns regarding applicability using the QUADAS-2 quality assessment tool (Figure 2). Up to 45% and 50% of the studies showed a high risk of bias regarding patient selection and index test, respectively. This was predominantly due to non-random patient selection and knowledge of reference standard results prior to interpreting the index test. Discussion Evolution of spectroscopy in clinical practice Since the first ex vivo RS study by Crow et al. (2004) using frozen tissue (Figure 3), significant technological advancements in optoelectronics, computationa l capacity, and machine-learning data analysis techniques have facilitated rapid and real-time applications[24]. In the first in vivo study, Draga et al. (2010) introduced a fiber optic probe with a 2.1-mm external diameter via a cystoscope to acquire RS measurements imme- diately before TURBT[19]. Their algorithm achieved a sensitivity of 85% and a modest specificity of 79% in distinguishing BCa from normal tissue, thus highlighting the challenges in clinical translation for RS. Subsequent RS studies have implemented hard- ware and software improvements to optimize diag- nostic accuracy. Barman et al. (2012) introduced a confocal fiber optic probe that limited sampling to 300 μm. FIGURE 3. Figures from the first ex vivo Raman Spectroscopy study by Crow et al. (2004) using frozen tissue. A) The mean Raman spectra measured for each of the pathological groups. B) Scatter plots of the scores of linear discriminant function 1 vs. 2 showing clustering in the eight-group algorithm. C) The prediction power of the eight-group diagnostic algorithm demonstrating a sensitivity and specificity of 93% and 98%, respectively. (Reproduced with permission.[24]) Adeno: adenocarcinoma; a.u.: absorption units; CIS: carcinoma in situ; Cyst: cystitis; ldf; linear discriminant function; Sq: squamous; TCC: transitional cell carcinoma. By suppressing spectral information from deeper tissue layers beyond the region of interest, diagnostic accuracy improved to 86% sensitivity and 100% specificity[25]. Following a similar principle of shallow tissue sampling, Stomp-Agenant et al. (2022) developed a superficial fiber optic probe with a measuring depth of 200 μm. This significantly reduced the signal-to-noise ratio compared to a regular probe, improving accuracy to 90% sensitiv- ity and 87% specificity[18]. In add it ion to t he ha rdwa re i mprovements described, advancements in computational capacity and machine-learning data analysis techniques continue to enhance the diagnostic accuracy of spectroscopy. Chen et al. (2018), analyzed 32 snap-frozen bladder specimens using a fiber optic probe, similar to previous studies, but combined PCA with artificial neural network (ANN) modelling to achieve a sensitivity of 98% and speci- ficity of 96% in detecting high-grade BCa. ANN is a powerful, self-adaptive, and data-driven pattern recog- nition method capable of capturing non-linear char- acteristics of the data[26]. After comparing ANN with other popular classifications methods such as linear discriminant analysis (LDA) or support vector machines (SVM) in dozens of studies, the authors noted that ANN constantly outperformed other techniques. While no single spectroscopy technique has proven to be perfect, a multimodal approach is likely to be required. RS has been combined with a concurrent diag- nostic method in 3 studies, including photodynamic diagnosis[19], optical coherence tomography[27], and f luorescence and diffuse ref lectance[28]. Although a 331330 SIUJ.ORG SIUJ • Volume 4, Number 4 • July 2023SIUJ • Volume 4, Number 4 • July 2023 SIUJ.ORG REVIEW Point-of-Care Diagnosis of Bladder Cancer With Vibrational Spectroscopy: A Systematic Review http://SIUJ.org http://SIUJ.org References 1. Richters A, Aben KKH, Kiemeney LALM. The global burden of urinary bladder cancer: an update. World J Urol.2020;38(8):1895 –1904. doi: 10.1007/s00345-019-0298 4 - 4. PMID: 31676912; PMCID: PMC7363726. 2. Sylvester RJ, van der Meijden APM, Oosterlinck W, et al. Predicting recurrence and progression in individual patients with stage Ta T1 bladder cancer using EORTC risk tables: a combined analysis of 2596 patients from seven EORTC trials. Eur Urol.2006;49(3):466– 477; discussion 475 – 477. doi: 10.1016/j.eururo.2005.12.031. PMID: 16442208. 3. Jocham D, Witjes F, Wagner S, Zeylemaker B, van Moorselaar J, Grimm MO, et al. Improved detection and treatment of bladder cancer using hexaminolevulinate imaging: a prospective, phase III multicenter study. J Urol.2005;174(3):862–866; discussion 866. doi: 10.1097/01. ju.0000169257.19841.2a. PMID: 16093971. 4. Schmidbauer J, Witjes F, Schmeller N, Donat R, Susani M, Marberger M; Hexvix PCB301/01 Study Group. Improved detection of urothelial carcinoma in situ with hexaminolevulinate fluorescence cystoscopy. J Urol.2004;171(1):135–138. doi: 10.1097/01.ju.0000100480.70769.0e. PMID: 14665861. 5. Cauberg ECC, de Bruin DM, Faber DJ, van Leeuwen TG, de la Rosette JJ, de Reijke TM. A new generation of optical diagnostics for bladder cancer: technology, diagnostic accuracy, and future applications. Eur Urol.2009;56(2):287–297. doi: 10.1016/j.eururo.2009.02.033. PMID: 19285787. 6. AA, Aboul-Enein HY. Vibrational spectroscopy applications in drugs analysis. In: Lindon JC, Tranter GE, Koppenaal DW, eds. Encyclopedia of Spectroscopy and Spectrometry. 3rd ed. Academic Press; 2017:575–581. 7. European Pharmaceutical Review. Pharmaceutical analysis with portable spectrometers. Published February 22, 2022. Available at: https://www.europeanpharmaceuticalreview.com/article/168732/ pharmaceutical-analysis-with-portable-spectrometers/. Accessed June 21, 2023. 8. Ly ng F M, Ramos IRM, Ibr ahim O, B y r ne H J. V ibr ational microspectroscopy for cancer screening. Appl Sci.2015;5(1):23–35. doi: 10.3390/app5010023. 9. Moisoiu V, Iancu SD, Stefancu A, Moisoiu T, Pardini B, Dragomir MP, et al. SERS liquid biopsy: an emerging tool for medical diagnosis. Colloids Surf B Biointerfaces.2021;208:112064. doi: 10.1016/j. colsurfb.2021.112064. PMID: 34517219. 10. Zacharovas E, Velička M, Platkevičius G, Čekauskas A, Želvys AN, Niaura G,et al. Toward a SERS diagnostic tool for discrimination between cancerous and normal bladder tissues via analysis of the extracellular fluid. ACS Omega.2022;7(12):10539–10549. doi: 10.1021/ acsomega.2c00058. PMID: 35382275; PMCID: PMC8973049. 11. Afara IO, Shaikh R, Nippolainen E, Querido W, Torniainen J, Sarin JK, et al. Characterization of connective tissues using near-infrared spectroscopy and imaging. Nat Protoc.2021;16(2):1297–1329. doi: 10.1038/s41596-020-00468-z. PMID: 33462441. 12. Beć KB, Huck CW. Break through potential in near-infrared spectroscopy: spectra simulation. A review of recent developments. Front Chem.2019;7:48. doi: 10.3389/fchem.2019.00048. PMID: 30854368; PMCID: PMC6396078. 13. Anderson DJ, Anderson RG, Moug SJ, Baker MJ. Liquid biopsy for cancer diagnosis using vibrational spectroscopy: systematic review. BJS Open.2020;4(4):554 –562. doi: 10.1002/bjs5.50289. PMID: 32424976; PMCID: PMC7397350. 14. Kallaway C, Almond LM, Barr H, Wood J, Hutchings J, Kendall C, et al. Advances in the clinical application of Raman spectroscopy for cancer diagnostics. Photodiagnosis Photodyn Ther.2013;10(3):207–219. doi:10.1016/j.pdpdt.2013.01.008. PMID: 23993846. 15. Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ.2021;372:n71. doi:10.1136/bmj. n71. PMID: 33782057; PMCID: PMC8005924. 16. Whiting PF, Rutjes AWS, Westwood ME, Mallett S, Deeks JJ, Reitsma JB, et al.; QUADAS-2 Group. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med.2011;155(8):529–536. doi: 10.7326/0003-4819-155-8- 201110180-00009. PMID: 22007046. 17. Ahmed EG, Al-Muslet NA, Ahmed MM, Moharam M, Musaad W. The use of fourier infrared spectroscopy and laser-Raman spectroscopy in bladder malignancy diagnosis, a comparative study. Appl Phys Res.2010;2(1):108–117. 18. Stomp-Agenant M, van Dijk T, R. Onur A, Grimbergen M, van Melick H, Jonges T, et al. In vivo Raman spectroscopy for bladder cancer detection using a superficial Raman probe compared to a nonsuperficial Raman probe. J Biophotonics.2022;15(6):e202100354. doi: 10.1002/jbio.202100354. PMID: 35233990. 19. Draga ROP, Grimbergen MCM, Vijverberg PLM, van Swol CF, Jonges TG, Kummer JA, et al. In vivo bladder cancer diagnosis by high- volume Raman spectroscopy. Anal Chem.2010;82(14):5993–5999. doi: 10.1021/ac100448p. PMID: 20524627. 20. Stone N, Hart Prieto MC, Crow P, Uff J, Ritchie AW. The use of Raman spectroscopy to provide an estimation of the gross biochemistry associated with urological pathologies. Anal Bioanal Chem.2007;387(5):1657–1668. doi: 10.1007/s00216-006-0937-9. PMID: 17123068. 21. Hughes C, Iqbal-Wahid J, Brown M, Shanks JH, Eustace A, Denley H, et al. FTIR microspectroscopy of selected rare diverse sub-variants of carcinoma of the urinary bladder. J Biophotonics.2013;6(1):73–87. doi: 10.1002/jbio.201200126. PMID: 23125109. 22. Pezzei C, Brunner A, Bonn GK, Huck CW. Fourier transform infrared imaging analysis in discrimination studies of bladder cancer. Analyst.2013;138(19):5719–5725. doi: 10.1039/c3an01101a. PMID: 23897512. Recent improvements in machine-learning analytical techniques have led to substantial progress in research and industry[11,12]. NIRS has the potential to provide real-time molecular data, analogous to handheld ultra- sound devices, with low computational requirements (6 Kb per spectrum), making it possible to be performed on mobile devices in line with the evolution toward ambulatory and personalized care[38]. Compared to the aforementioned techniques, NIRS spectra can be obtained from greater sample thickness, allowing for easier sample handling, and it is fast without the need for a laser, unlike RS. Additionally, near-infra- red light penetrates deeper into human tissues, caus- ing less photodamage and safer tissue probing[39]. Considering that NIRS and RS provide complementary information when analyzing the same sample, combin- ing the 2 techniques in a multimodal approach could potentially further enhance diagnostic accuracy. With ongoing advancements in spectroscopy tech- nology, machine-learning analytical techniques, and multimodal approaches to improve accuracy, VS offers several potential advantages over standard histopathol- ogy: rapid, label-free, and operator independent. When used in conjunction with fiber optic probes in endoscopy, it may help reduce cases of incomplete tumour resection and lower the risk for recurrence. It can serve as a tool to aid clinical decision-making in real time, providing a quick and safe assessment of stage and grade, allowing urologists to reduce over- or under-treatment of BCa. In an increasingly frail population with rising antico- agulant use, these improvements could reduce adverse events related to surgery and expedite the staging and grading of urothelial cancer of the bladder. Conclusions Although VS is a mature technology in analytical chemistry, its use in medical diagnostics is still in its infancy. Recent advances in technology and computing power and reductions in equipment costs and size have facilitated a shift in focus from the laboratory to the bedside. As fiber optic probes for spectroscopy become commercially available, their use in combination with a conventional cystoscope opens up the exciting possibility for real-time diagnostic imaging of BCa. RS has demonstrated high levels of diagnostic accuracy, which continue to improve with advancements in SERS. However, studies are small and highly heterogeneous. Larger spectroscopy studies with robust reporting methods and a multimodal approach are needed to assess not only the overall diagnostic accuracies but also the optimal utilization of this emerging technology. Ongoing research into modalities such as SERS and NIRS holds great promise, making spectroscopy an exciting and dynamic field in urological diagnostics with the potential to enhance intraoperative decision- making. 333332 SIUJ.ORG SIUJ • Volume 4, Number 4 • July 2023SIUJ • Volume 4, Number 4 • July 2023 SIUJ.ORG REVIEW Point-of-Care Diagnosis of Bladder Cancer With Vibrational Spectroscopy: A Systematic Review http://SIUJ.org http://SIUJ.org 23. Jin D, Wang X, Fu B, Li T, Chen N, Chen Z, et al. Raman spectroscopy of luminal subtype and basal subtype muscle invasive bladder cancer. Int J Clin Exp Med.2019;12(5):5447–5453. 24. Crow P, Uff JS, Farmer JA, Wright MP, Stone N. The use of Raman spectroscopy to identify and characterize transitional cell carcinoma in vitro. BJU Int.200 4;93(9):1232–1236. doi: 10.1111/j.14 6 4 - 410X.2004.04852.x. PMID: 15180613. 25. Barman I, Dingari NC, Singh GP, Kumar R, Lang S, Nabi G. Selective sampling using confocal Raman spectroscopy provides enhanced specificit y for urinar y bladder cancer diagnosis. Anal Bioanal Chem.2012;404(10):3091–3099. doi: 10.1007/s00216-012-6424-6. PMID: 23052868. 26. Chen H, Li X, Broderick N, Liu Y, Zhou Y, Han J, et al. Identification and characterization of bladder cancer by low-resolution fiber-optic Raman spectroscopy. J Biophotonics.2018;11(9):e201800016. doi: 10.1002/ jbio.201800016. PMID: 29797794. 27. Placzek F, Cordero Bautista E, Kretschmer S, Wurster LM, Knorr F, González-Cerdas G, et al. Morpho-molecular ex vivo detection and grading of non-muscle-invasive bladder cancer using forward imaging probe based multimodal optical coherence tomography and Raman spectroscopy. Analyst.2020;145(4):1445–1456. doi: 10.1039/ c9an01911a. PMID: 31867582. 28. Morselli S, Baria E, Cicchi R, Liaci A, Sebastianelli A, Nesi G, et al. The feasibility of multimodal fiber optic spectroscopy analysis in bladder cancer detection, grading, and staging. Urologia.2021;88(4):306–314. doi: 10.1177/03915603211007018. PMID: 33789562. 29. Jin H, Lin T, Han P, Yao Y, Zheng D, Hao J, et al. Efficacy of Raman spectroscopy in the diagnosis of bladder cancer: a systematic review and meta-analysis. Medicine (Baltimore).2019;98(47):e18066. doi: 10.1097/MD.0000000000018066. PMID: 31764837; PMCID: PMC6882629. 30. Huttanus HM, Vu T, Guruli G, Tracey A, Carswell W, Said N, et al. Raman chemometric urinalysis (Rametrix) as a screen for bladder cancer. PLoS One.2020;15(8):e0237070. doi: 10.1371/journal.pone.0237070. PMID: 32822394; PMCID: PMC7446794. 31. Liu Z, Zhang P, Wang H, Zheng B, Sun L, Zhang D, et al. Raman spectrum- based diagnosis strategy for bladder tumor. Urol Int.2022;106(2):109– 115. doi: 10.1159/000518877. PMID: 34515249. 32. Hu D, Xu X, Zhao Z, Li C, Tian Y, Liu Q, et al. Detecting urine metabolites of bladder cancer by sur face-enhanced Raman spectroscopy. Spectrochim Acta A Mol Biomol Spectrosc.2021;247:119108. doi: 10.1016/j.saa.2020.119108. PMID: 33161263. 33. Gok S, Aydin OZ, Sural YS, Zorlu F, Bayol U, Severcan F. Bladder cancer diagnosis from bladder wash by Fourier transform infrared spectroscopy as a novel test for tumor recurrence. J Biophotonics.2016;9(9):967–975. doi: 10.1002/jbio.201500322. PMID: 27041149. 34. Kuepper C, Kallenbach-Thieltges A, Juet te H, Tannapfel A, Großerueschkamp F, Gerwert K. Quantum cascade laser-based infrared microscopy for label-free and automated cancer classification in tissue sections. Sci Rep.2018;8(1):7717. doi: 10.1038/s41598-018- 26098-w. PMID: 29769696; PMCID: PMC5955970. 35. Cordero E, Rüger J, Marti D, Mondol AS, Hasselager T, Mogensen K, et al. Bladder tissue characterization using probe-based Raman spectroscopy: evaluation of tissue heterogeneity and influence on the model prediction. J Biophotonics.2020;13(2):e201960025. doi: 10.1002/jbio.201960025. PMID: 31617683; PMCID: PMC7065650. 36. Ali JH, Wang WB, Zevallos M, Alfano RR. Near infrared spectroscopy and imaging to probe differences in water content in normal and cancer human prostate tissues. Technol Cancer Res Treat.2004;3(5):491–497. doi: 10.1177/153303460400300510. PMID: 15453814. 37. Gu Y, Chen WR, Xia M, Jeong SW, Liu H. Effect of photothermal therapy on breast tumor vascular contents: noninvasive monitoring by near- infrared spectroscopy. Photochem Photobiol.2005;81(4):1002–1009. doi: 10.1562/2004-09-05-RA-305. PMID: 15807632. 38. Adegoke JA, Gassner C, Sharma VJ, Patel SK, Jackett L, Afara IO, et al. Near-infrared spectroscopic characterization of cardiac and renal fibrosis in fixed and fresh rat tissue. Anal Sens.2023;3(1):e202200030. doi: 10.1002/anse.202200030. 39. Wu S, But t H-J. Near-infrared-sensitive materials based on upconverting nanoparticles. Adv Mater.2016;28(6):1208–1226. doi: 10.1002/adma.201502843. PMID: 26389516. 40. Crow P, Molckovsky A, Stone N, Uff J, Wilson B, WongKeeSong L M . A s s e s s m e n t o f f i b e r o p t i c n e a r- i n f r a r e d r a m a n spectroscopy for diagnosis of bladder and prostate cancer. Urology.2005;65(6):1126–1130. doi: 10.1016/j.urology.2004.12.058. PMID: 15913721. 41. de Jong BWD, Bakker Schut TC, Maquelin K, van der Kwast T, Bangma CH, Kok DJ, et al. Discrimination between nontumor bladder tissue and tumor by Raman spectroscopy. Anal Chem.2006;78(22):7761–7769. doi: 10.1021/ac061417b. PMID: 17105169. 42. Al-Muslet NA, Ali EE. Spectroscopic analysis of bladder cancer tissues using Fourier transform infrared spectroscopy. J Appl Spectrosc.2012;79(1):139–142. 43. Pavlov V, Bilyalov A, Gilmanova R, Yakupov RR, et al. The use of intelligent data processing techniques of Raman spectroscopy for the diagnosis of malignant tumors. Bashkortostan Med J.2016;13. 44. Yousif ES, Abdulkareem DT, Enad alboaisa Ns, Mohammad EJ. Detection of urinary bladder cancer by (ATR-F TIR) spectroscopy. System Rev Pharm.2020;11(12):1932–1937. 45. Taieb A, Berkovic G, Haifler M, Cheshnovsk y O, Shaked NT. Classification of tissue biopsies by Raman spectroscopy guided by quantitative phase imaging and its application to bladder cancer. J Biophotonics.2022;15(8):e202200009. doi: 10.1002/jbio.202200009. PMID: 354887 334 SIUJ • Volume 4, Number 4 • July 2023 SIUJ.ORG REVIEW http://SIUJ.org