**2. Results**

In total, 220 patients with IC/BPS (180 women, and 40 men) who had received BoNT-A injections were enrolled. The mean age and IC/BPS duration were 54.1 ± 13.4 and 13.8 ± 10.1 years, respectively. The mean maximal bladder capacity (MBC) under anesthesia was 652 ± 204 mL; with 76 (35%), 114 (51.4%), and 20 (9.1%) patients having grade I glomerulation, grade II–III glomerulation, and HIC, respectively. The mean IC symptom index (ICSI) and mean IC problem index (ICPI) were 12.7 ± 3.7 and 12.0 ± 3.3 points, respectively. The mean numerical rating scale (NRS) score for bladder pain was 5.1 ± 2.7 points. In total, 124 (56%) patients had voiding dysfunction, which included bladder neck dysfunction (*n* = 8, 4%), dysfunctional voiding (*n* = 14, 6%), and poor external sphincter relaxation (102, 46%) whilst under videourodynamic study (VUDS), and all patients had storage dysfunction, which included bladder hypersensitivity (*n* = 198, 90%) and detrusor overactivity (*n* = 22, 10%); 214 (97%) patients had bladder pain or an intense urge to void during the potassium chloride (KCl) infusion test. In total, 180 (82%) patients reported having bladder, pelvic, or lower abdominal pain or discomfort (Table 1).


**Table 1.** IC/BPS patients' characteristics and VUDS parameters according to the treatment outcome (*n =* 220).

\* Significant *p* < 0.05.

Among the 220 patients receiving BoNT-A injections, 133 (60%) reported unsatisfactory treatment outcomes which were evaluated using the global response assessment (GRA) at 6 months later, and 87 (40%) reported satisfactory treatment outcomes. The unsatisfactory group had a significantly higher rate of urinary tract symptoms (LUTS), including urinary frequency (88.9% vs. 69.9%), nocturia (87.5% vs. 74.2%), and urinary retention (6.9% vs. 1.1%) than the satisfactory group. However, the unsatisfactory group had lower urge incontinence (8.3% vs. 18.3%) rate. Nine (10%) patients in the satisfactory group underwent electrical coagulation of Hunner's lesion. Moreover, as shown in Table 1, there were no significant differences in the subjective awareness items which included the ICSI and ICPI of the O'Leary–Sant symptom scale (OSS), and the NRS for bladder pain severity between the two groups, regardless of their self-reported treatment outcomes. However, the satisfactory group had a much larger voided volume (273 ± 130 vs. 236 ± 111 mL, *p* = 0.026) at the baseline. Moreover, there were no significant differences in age, sex, IC duration, pain severity, IC symptoms, problem severity, VUDS parameters, glomerulation grade, MBC, and KCl test results at the baseline between the two groups (Table 1).

In contrast, according to their bladder condition, we referred to a previous study that used statistical analysis of the receiver operating characteristic curve to predict satisfactory outcomes [11] and to further define both the MBCs greater or less than 760 mL and the phenotype divided on bladder capacity combined with the glomerulation grade. The unsatisfactory group was found to have a significantly smaller bladder capacity of <760 mL in comparison to the satisfactory group (*n* = 103, 77% vs. *n* = 53, 61%, *p* = 0.007). Interestingly, among 220 patients, excluding those with HIC, who had their urine collected for urinary biomarker analysis, the levels of the inflammatory biomarker monocyte chemoattractant protein-1 (MCP-1) and the oxidative stress biomarker 8-isoprostane were significantly higher in the unsatisfactory group (379 ± 517 vs. 229 ± 259, *p* = 0.031; 115 ± 245 vs. 44.8 ± 48.1, *p* = 0.017) (Table 2). Among the patients with satisfactory outcomes, 37% (*n* = 32) received intravesical BoNT-A injections more than four times, whereas, in the unsatisfactory group, only 27% (*n* = 36) received the BoNT-A injection more than four times.

**Table 2.** Bladder condition and urinary biomarkers of the IC/BPS patients stratified according to the treatment outcome (*n =* 220).


MBC: Maximal bladder capacity, IL-8: interleukin-8, CXCL10: C-X-C motif chemokine ligand 10, MCP-1: monocyte chemoattractant protein-1, BDNF: brain-derived neurotrophic factor, IL-6: Interleukin 6, RANTES: regulated upon activation/normal T cell expressed and secreted, PGE2: prostaglandin E2, TNF-α: tumor necrosis factor-alpha, 8-OHdG: 8-hydroxy-2-deoxyguanosine, TAC: total antioxidant capacity; \* significant *p* < 0.05.

Finally, to determine the IC/BPS patients' subjective or objective influencing factors, including age, sex, IC duration, IC symptoms and problem severity, VUDS parameters, and phenotype, a supervised machine learning algorithm was used to predict the probability of an outcome. The results of the logistic regression indicated that IC/BPS symptoms and problem severity (odds ratio: 1.06, 95% CI: 1.01–1.13, *p* = 0.031) and detrusor pressure were significant predictors of treatment outcomes. Patients with grade 2–3 glomerulation and MBC < 760 mL might have poor treatment outcomes compared to those with grade 0 to 1 glomerulation and MBC ≥ 760 mL (odds ratio: 0.04, 95% CI: 0.17–0.97, *p* = 0.042) (Figure 1). However, adverse events including hematuria after injection in 6 (2.7%) patients, 4 (1.8%) urinary tract infections, and 36 (16%) mild dysuria cases were found, but these were without reports of urinary retention from any of the patients treated.

**Figure 1.** The forest plot results from the logistic regression for predicting satisfactory treatment outcome (GRA ≥ 2) effect factors. \* significant *p* < 0.05.
