The clinical study of preoperative normalized apparent diffusion coefficient of 3.0-T MRI in predicting muscular invasion of bladder cancer.

BACKGROUND

To investigate the clinical application value of preoperative normalized apparent diffusion coefficient (nADC) from 3.0-T magnetic resonance imaging (MRI) in predicting muscular invasion of bladder cancer (BC).

METHODS

This retrospective study enrolled BC patients confirmed by surgical pathology at a tertiary hospital between November 2019 and May 2022. All patients underwent preoperative pelvic MRI with two-orientation diffusion-weighted imaging: axial image was acquired routinely, followed by sagittal or coronal image to display the tumor base optimally (selecting based on three-orientation T2-weighted imaging). Patients were divided into muscle-invasive BC (MIBC) and non-muscle-invasive BC (NMIBC) groups based on histopathology. Tumor ADC normalized to urine as reference. Two radiologists who were double-blind, measured the two-orientation ADC values of tumors and bladder urine on the same slice three times respectively, and calculated the average as ADCtumor value and ADCurine value. One of the radiologists repeated the measurements after 6 months. NADC = ADCtumor value / ADCurine value. Intraclass correlation coefficient (ICC) was used to assess inter- and intra-reader consistency, with ICC ≥ 0.75 indicating good consistency. With histopathology as the reference standard, receiver operating characteristic curves were used to compared the diagnostic performance of nADC and ADCtumor value, and to determine cut-off value, sensitivity, and specificity using the maximum Youden index.

RESULTS

The study included 109 patients (56 MIBC and 53 NMIBC). The inter- and intra-reader consistency for the two readers in measuring tumor and urine ADC values were good (ICC > 0.75, p < 0.001). Both nADC and ADCtumor value were negatively correlated with tumor muscular invasion (rnADC = -0.723, rADC = -0.600, p < 0.001). MIBC showed lower nADC (0.292 vs. 0.375, p < 0.001) and ADCtumor value (0.894 × 10− 3mm2/s vs. 1.107 × 10− 3mm2/s, p < 0.001) than NMIBC. The best cut-off values for predicting muscular invasion were 0.334 for nADC and 0.985 × 10− 3mm2/s for ADCtumor value. The area under the curve (AUC), sensitivity and specificity of nADC (0.917, 89.3% and 83.0%) were higher than those of ADCtumor value (0.847, 75.0% and 77.4%), and the difference in AUC was statistically significant (p = 0.002).

CONCLUSION

The preoperative nADC and ADCtumor value are negatively correlated with muscular invasion in BC. The lower the nADC and ADCtumor value, the more likely the tumor invades muscle layer. NADC demonstrates superior diagnostic performance compared to ADCtumor value, providing more valuable preoperative assessment of BC invasiveness to guide personalized therapeutic decisions.