BUC is the most common malignant tumor of the bladder, characterized by strong invasiveness that facilitates infiltration into muscular layer and invade surrounding tissues. This makes complete removal the tumor via TURBT challenging, resulting in a significantly higher recurrence rate and lower survival rate for patients [12]. IPB is the most common benign tumor of the bladder, generally showing limited growth and not invading the muscular layer [13]. However, due to its unclear pathogenesis and low incidence rate, clinical symptoms and imaging manifestations often overlap with BUC, many radiologists lack sufficient experience in diagnosing it accurately unless peripheral abnormalities or signs of metastasis are observed, which is not conducive to guiding treatment and assessing patient prognosis. In this study, we examined the disparities in multiple qualitative and quantitative indicators present in CTU images between BUC and IPB, and finally searched for 4 easy-to-use quantifiable features, all of which demonstrated satisfactory diagnostic efficacy. The nomogram, constructed by the combination of 4 independent risk factors, further improved the discrimination between BUC and IPB, achieving the highest AUC and specificity. This suggests that CTU-based quantitative features may aid improve the accuracy in distinguishing between BUC and IPB preoperatively, thereby contributing to enhancing the reliability of imaging diagnostic reports.

In this study, there were no significant inter-group differences observed in terms of gender, morphology, number of lesion, calcification, cystic degeneration, or plain CT values, indicating that distinguishing BUC from IPB based solely on basic qualitative characteristics observed by the naked eye is difficult. Both of them are prevalent among men and way present with calcification, cystic degeneration, and multiple lesions are not uncommon, which appear as limited bladder wall thickening or soft tissue nodules on non-enhanced CT [14, 15]. In clinical practice, both BUC and IPB are generally considered to be prevalent in middle-aged and elderly. In all participants enrolled in this study, the mean age of BUC was higher than the that of IPB (63.51±12.23 years vs. 55.00±11.80 years, P=0.003), suggesting that age may serve as a potential predictive factor for differentiating them. However, the multivariate analysis failed to identify age as an independent risk factor. We postulated there were two reasons: first, the sample size was small, particularly in IPB with only 25 cases, which was insufficient to fully reflect age differences. Second, there may have been selection bias as this study included BUC that did not invade structures outside the bladder, with slower cell proliferation and smaller overall lesion size, resulting in late onset of clinical symptoms and delaying detection. However, these speculations need validation through larger-scale studies with more comprehensive staging of BUC.

Among the 4 independent risk factors, the junction smoothness was the only subjective qualitative indicator, which indirectly reflects image features around tumors. The inter-observer agreement between two radiologists' measurements was high with an ICC of 0.841. 65.4% (51/78) BUC had junction abnormalities, whereas only 32.0% (8/25) IPB caused non-smoothness at the junction. This may be attributed to their biological behavior. The tumor cells of BUC grow at different rates in all directions, resulting in irregular or lobulated tumor shapes, which promotes the precipitation and adsorption of urate crystals around the focus [16]. Moreover, about 25% of BUC infiltrates downward into the muscular layer, causing thickening and stiffening of the basal bladder wall, while the adjacent normal bladder wall was soft and mobile, creating a junction bulge or disruption of continuity [17]. IPB exhibits slow growth and regular morphology, with its surface being covered by normal urinary tract epithelium and without involving the muscle layer. There, there is no significant difference between the lesion margin and the adjacent bladder wall [18].

The maximum longitudinal diameter and tumor-wall interface quantitatively reflect the lesion size. The findings of this study showed an inverse correlation between tumor height and the likelihood of presenting with BUC. IPB typically appear as narrow basal strips or sphere-like shapes on CT images, with a longitudinal diameter similar to the transverse diameter. Previous research have shown that tipped is a specific presentation of IPB [14, 19, 20]. In contrast, the transverse diameter of BUC was significantly larger than its longitudinal diameter, indicating that the tumor show more of a broad base and the growth direction spread mainly to the bladder wall on both sides. The author posits that the aspect ratio may be a superior indicator for evaluating morphological characteristics of tumors compared to basal and height measurements, even if it was not included as a reference standard in the nomogram. For example, the aspect ratio is frequently utilized in predicting BUC staging, and one study have shown that when the aspect ratio is <0.605, it suggests a heightened likelihood of myofilament infiltration [21]. Tumor-wall interface, when used along, is the most effective independent risk factor for identifying BUC and IPB. There was no statistically different from the AUC of nomogram. This easily observable and measurable metric performed nearly perfectly in terms of inter-reader agreement. Combined use with other indicators in a multivariate logistic regression equation significantly improved the predictive efficacy, suggesting that other indicators are complementary to tumor-wall interface in predicting tumor benignity and malignancy. Previous studies have shown that the tumor-wall interface of lesions without stick was wider than that of tipped lesions, and the risk of myometrial invasion is higher [22, 23]. Ahn et al. measured the length of tumor-bladder contact on T2WI, DWI and ADC images in combination with the Vesical Imaging-Reporting and Data System (VI-RADS) to predict muscular invasion of BUC, and showed that the tumor-wall interface was an independent risk factor (OR range 1.9-2.0) for differentiating MIBC from NMIBC, compensating for the lack of VI-RADS as a qualitative indicator [24]. A variety of derived quantitative indicators have been used for staging other tumors that were similar to tumor-wall interface. For example, the extent of contact with the pleura has been used as a criterion for judging pleural invasion in lung cancer [25]. Similarly, the length of contact with the envelope has been used to predict extracapsular extension of prostate cancer [26].

Since tumor growth depends is dependent on angiogenesis, varying degrees of neovascularization and malformation can reflect different levels of tumor differentiation [27]. The presence or absence of arteriovenous fistula, vascular tortuosity, distribution, vessel size and interstitial edema can affect the perfusion of contrast medium within the lesion. The arterial reinforcement rate represents the blood supply characteristics of the lesion. The results demonstrated that the arterial reinforcement rate of BUC was significantly higher than that of IPB, indicating that the blood supply of BUC was richer than that of IPB, and the heterogeneity of tumor vessels was higher, which is consistent with the pathological manifestations [28, 29]. The hyperplastic epithelial nests of IPB are parallel arranged epithelial cells with minimal fibrovascular structures, and derive their blood supply primarily from the surrounding vasculatures of the tumor cell mass. By comparison, BUC is a malignant neoplasm characterized by extensive neovascularization and rapid growth, and most vessels extend to the interior of the tumor in the shape of dendrites at the base of the tumor. Therefore, typical BUC shows obvious enhancement in arterial phase, with maximum enhancement in venous phase and diminished enhancement in delayed phase. Conversely, the typical IPB shows moderate to severe delayed intensification. When describing the enhancement characteristics, we only analyzed the difference between arterial phase CT and non-enhanced CT value, but did not measure venous phase and delayed CT value. The reason is that the dynamic changes of the enhancement degree of lesions analyzed by simultaneous measurement of arterial and venous CT values are closely related to contrast medium injection time and individual differences such as cycle period, which may increase the error of measurement results. In addition, the contrast medium was excreted into the bladder through the urinary tract during the delayed phase, resulting in a "jet" image that interfered with accurate CT value measurement of adjacent lesions.

At present, routine examinations for the diagnosis of bladder tumors include ultrasound, CT, MRI and cystoscopy. However, due to the low soft tissue separation rate and limited scanning range of ultrasound, it is difficult to make qualitative diagnosis only according to sonogram. MRI offers clear advantages in evaluating myometrial invasion and lymph node metastasis [30], its imaging time is lengthy and image quality can be affected by intestinal peristalsis. Furthermore, the conventional scan image is thicker, and the display effect of micro-lesion and calcification is not distinct. Cystoscopy is subject to operator-dependent variability and incomplete sampling, which may lead to inaccurate pathological results. In addition, the possibility of damaging the urethra and causing urinary tract infection limits its frequent use in diagnosing and postoperative monitoring bladder tumors. In this study, CTU was used to identify BUC and IPB because of its 5 major advantages: (1) non-invasive, fast imaging and economical cost; (2) relatively objective data acquisition that is not influenced by operator level; (3) the target vessels are reconstructed using MIP mode, which can show the vascular structures within and around the lesion; (4) the excretion phase images provided clearer lesion morphology and contour details, as well as more precise diameter measurements; (5) the ability to quantitatively measure multiple features of the lesion.

The human papillomavirus (HPV) hinders the function of tumor suppressor proteins by means of its viral oncogenic proteins E6 and E7, disrupting the regulation of cell cycle and DNA repair mechanisms. Consequently, this impairs genomic stability in cells, thereby elevating the risk of cellular malignancy and tumor recurrence. The study conducted by Sarier et al. [31] involved a 2-year regular follow-up of patients with BUC, revealing that the recurrence rate was higher in HPV-positive patients (47.3%) compared to HPV-negative patients (36.8%). These findings suggest that HPV infection could potentially serve as a valuable tool for distinguishing BUC from IPB prior to surgery and assessing tumor prognosis, warranting further investigation.

Several limitations of this study are noteworthy. First, this was a single-center retrospective study, and selection bias is inevitable. Second, the sample size was small and unbalanced; therefore further studies require an increased sample size. Third, the assessment involved only a limited number of readers and did not analyze the intra-reader reproducibility of quantitative indicators. Finally, the index lesions were classified according to lesion size, which means that they may not necessarily by the highest grade lesions pathologically and could have caused some misleading results. A follow-up study will stratify the different grades of BUC and explore the differences in imaging performance with IPB.