Narrow-band Imaging Diagnostic Classification System for Bladder Tumors

1. Background： Bladder cancer is the most common malignancy of the urinary system in China, and its incidence is increasing year by year. Non-muscle invasive bladder cancer (NMIBC) can mostly be removed in a minimally invasive manner by transurethral resection of bladder tumor (TURBT), which has less impact on the patient's quality of life. In contrast, if the tumor progresses to muscle-invasive bladder cancer, radical cystectomy with urinary diversion is required, and patients need to carry a stoma bag for life, which significantly affects their quality of life. Early detection and early treatment can significantly improve the prognosis of bladder cancer patients. At present, cystoscopy under ordinary white light (WLC) is the main diagnostic method to check early bladder tumor and the presence of recurrence, and it is widely used in clinical practice, but its accuracy still needs to be improved. Therefore, in order to further improve the sensitivity and accuracy of cystoscopy, we need some optical imaging techniques as an adjunct to WLC to improve the visualization of tumors by contrast enhancement.
2. Narrow-band Imaging (NBI)： NBI is a diagnostic endoscopic illumination technology developed by Olympus corporation and integrated into the camera of the cystoscope. The NBI device filters out the red spectrum from white light, allowing the blue and green spectra to be absorbed by hemoglobin. The different penetration of the two spectra within the bladder mucosa enhances the contrast between vascular tissue and normal bladder mucosa, improving the visualization of tumor neovascularization. The shorter wavelength blue light (415 nm) band enters the superficial layer of the mucosa, making the superficial capillaries brownish-black, and the longer wavelength green light (540 nm) band enters the deeper layer of the mucosa, making the submucosal vessels green. Thus, papillary tumors exhibit brownish-green lesions due to the enhancement of submucosal capillaries and dark brown lesions due to the enhancement of the dense superficial vascular system.
3. NBI classification criteria for different tumors： Currently, NBI classification criteria have been established for both laryngeal tumors and intestinal tumors. Clinical practice has confirmed that the sensitivity and specificity of NBI in the diagnosis of laryngeal mucosal lesions are higher than those of WLC. researchers at the Cancer Hospital of the Chinese Academy of Medical Sciences have classified laryngeal lesions into 5 categories from precancerous to cancerous lesions and confirmed through clinical validation that the classification criteria can help identify precancerous lesions and tumors in the larynx at an early stage. The Japanese NBI expert group also proposed a general NBI classification for colorectal tumors in 2014, which classified colorectal lesions into polyps, low-grade intramucosal neoplasia, high-grade intramucosal neoplasia/ superficial mucosal invasive carcinoma, and deep mucosal invasive carcinoma according to their different morphological characteristics under NBI. Although the NBI technique has been clinically applied for more than 20 years, there is no uniform classification standard for the manifestation of different bladder tumors under the NBI pattern, and the microscopic diagnosis mostly relies on the clinician's experience, with the possibility of misdiagnosis and missed diagnosis.
4. Purpose of the study： This study plans to develop the NBI classification criteria for bladder lesions by summarizing and analyzing the morphological characteristics of different bladder lesions in the NBI pattern, using the pathological findings as a diagnostic basis, and assessing their sensitivity and specificity in order to guide clinical work for a more standardized and accurate diagnosis of bladder tumors and early recurrence, thus improving patient prognosis.
5. Research content： This is a prospective multicenter study, which includes all patients who need to undergo cystoscopy and take biopsies in our center and collaborating units, records the characteristics of performance under ordinary white light and NBI in different parts of the bladder during cystoscopy, analyzes the characteristics of NBI corresponding to different histopathology as the diagnostic criteria, establishes the diagnostic classification standard of bladder tumor NBI, and then verifies the reliability of this classification standard through clinical validation. There are two key issues to be addressed in this project: (1) to clarify the characteristics of different bladder lesions under NBI and establish a bladder tumor NBI diagnostic classification system based on the pathological results; (2) to clinically validate the accuracy of the classification criteria.
6. Cystoscopy method： The patient was placed in a lithotomy position, and the surgical area was disinfected with iodophor and covered with towels, and the urethral mucosal surface was anesthetized with 2% lidocaine gel. The morphological characteristics of the mucosa and blood vessels under the microscope were recorded, and tissue biopsy was performed in the suspected lesion area, and the biopsy tissue was stored in 10% formalin solution for pathological examination. All operations were performed by urologists of equal seniority who had undergone systematic training.
7. Sample size and establishment of classification method： Starting from January 1, 2021, cases were enrolled in our center and collaborating institutions simultaneously according to the inclusion and exclusion criteria. Because the different characteristics of mucosa and submucosal vessels of different nature of tumors in the NBI pattern were easier to distinguish, the proposed sample size was included in combination with the sample size included in the existing classification criteria of colorectal tumors and laryngeal tumors (n=139, 104) as follows: 100 cases in the no lesion group, 200 cases in the benign lesion group, and 200 cases in the malignant lesion group, and cystoscopy was performed according to the above method. For patients who did not require surgical treatment, the histopathological results of biopsy were used as the diagnostic criteria; for patients whose biopsy pathology suggested malignant bladder tumors and required TURBT or total cystectomy, the postoperative pathological results were used as the diagnostic criteria. According to the pathological results, all patients were divided into three groups: no lesion group, benign lesion group, and malignant lesion group. All pathological results were reported and reviewed by the hospital pathology center. The International Society of Urological Pathology (ISUP) classification standard was used for pathological classification, the WHO 2004 bladder cancer grading system was used for histological grading, and the Union International Control Cancer (UICC) 2017 TNM staging system was used for bladder cancer staging.
8. Blind design： The cystoscopy operation was performed by four urologists at each center who were trained and dedicated to the operation, separate from the primary clinician; neither the subject nor the investigator knew the patient number information before complete subject screening and enrollment, and the cystoscopy operator did not know the primary history information and relied only on the microscopic presentation to give the initial diagnosis; blinded data review was used.
9. Establishment of NBI diagnostic classification system for bladder tumors： Using the pathological results as the "gold standard", the patients were divided into no lesion group, benign lesion group and malignant lesion group, and analyzed the morphological characteristics of normal mucosa, benign and malignant lesions, focal mucosa and submucosal vessels, respectively, against the cystoscopic WLC and NBI microscopic manifestations, to summarize and establish the diagnostic classification system of bladder tumor NBI.
10. Clinical validation of the accuracy of the NBI diagnostic classification criteria for bladder tumors： Sample size calculation: There is no diagnostic classification standard for bladder tumor NBI, referring to the classification standard for colorectal tumor NBI, the diagnostic sensitivity of its lesionless group, benign lesion group and malignant lesion group are 75%, 91%, 35%, the specificity is 96%, 70%, 100%, also the tolerance error are 0.1, requiring one-sided test, α is 0.025, the certainty is 80%. In this phase, it is proposed to initially include a total of 564 patients in accordance with this sample size, according to the inclusion and exclusion criteria, prospectively, combined with the loss of follow-up rate (20%) in each group, according to which the 3 groups need to include at least 211, 127, and 226 study subjects, respectively. After the completion of the first part of the trial, the sample size can be corrected according to the established diagnostic classification criteria for NBI, after the initial estimation of its sensitivity and specificity.

The patients enrolled in the study were analyzed for cystoscopic NBI manifestations and given a preliminary diagnosis according to the diagnostic classification of bladder tumor NBI, also using tissue biopsy or surgical pathology as the "gold standard", and the accuracy of this classification was calculated.

Main evaluation indexes: Sensitivity, Specificity, Area under the ROC curve. Secondary evaluation metrics: Positive Predictive Value (PPV), Negative Predictive Value (NPV), Positive Likelihood Ratio (PLR), Negative Likelihood Ratio (NLR), Accuracy Likelihood Ratio (NLR), Accuracy, and Jouden Index.