ABSTRACT
This study aims to explore the possibility and bottleneck of clinical translation for an artificial intelligence (AI) diagnosis system for bladder cancer based on cystoscopy.We retrospectively collected videos of 101 bladder cancer patients from January to November 2023, at Sun Yat-sen Memorial Hospital, Sun Yat-sen University. Among these patients, with a median age of 63 years and 81.0% were male. The bladder cancer AI diagnosis system was utilized for diagnosis, and the accuracy of diagnoses from the videos was assessed. Additionally, a surgical evaluation scale was formulated to evaluate the quality of the videos, simulating clinical usage.The final test results showed a system sensitivity of 97.8%, a positive predictive value of 81.7%, specificity of 54.2%, and a negative predictive value of 92.3%. Furthermore, the surgical evaluation scale scores ranged from 3.96 to 4.69, indicating the feasibility of clinical application for this system.This study further quantitatively validated the accuracy of an artificial intelligence system using cystoscopy videos and assessed the potential for clinical application.
Subject(s)
Artificial Intelligence , Cystoscopy , Urinary Bladder Neoplasms , Humans , Urinary Bladder Neoplasms/diagnosis , Cystoscopy/methods , Retrospective Studies , Male , Middle Aged , Female , Sensitivity and SpecificityABSTRACT
Objective: To explore the value of ultra-short echo time (UTE)-T2* component analysis techniques in dynamic monitoring the morphological and biochemical changes in amateur marathon athletes' achilles tendon before and after the marathon. Methods: Twenty-nine amateur marathon runners were recruited between October 2020 and March 2021 in Zhuhai City, Guangdong Province, including 25 males and 4 females, aged from 24 to 50 (40±6) years old. All volunteers underwent bilateral achilles tendon MRI examination 1 week before the marathon, 48 hours after the race, and 1 month after the race. The shape and signal of the achilles tendon were evaluated by routine T1-weighted, proton density weighted with fat saturation sequence and different echo time (TE) UTE sequence, and the changes of achilles tendon after running was quantitatively analyzed by UTE-T2* sequence. The values of single-component analysis (T2*M), short T2* components (T2*S), and long T2* components (T2*L) and Fraction values were obtained using UTE-T2* sequence. The value of the whole achilles tendon was measured on the sagittal images of achilles tendon, and the Achilles tendon was equally divided into three subregions [muscle-tendon junction (MTJ), middle (MID), and insertion (INS)]. The region of interest was delineated by two radiologists independently. The intra-group correlation coefficient (ICC) was used to evaluate the consistency of the data measured by two radiologists. Nonparametric Friedman M test was used to compare the differences of T2*M, T2*S, T2*L and Fraction values in different time points and different subregions. Wilcoxon rank-sum test was used to compare the difference between 48 h post-race and pre-race T2*S values (ΔT2*S) of different distance, different running posture, different pace and different amount of training, in which ΔT2*S equals the T2*S value of 48 h post-race minus the T2*S value of pre-race. Results: On the sequence of short TE (TE≤0.6 ms), achilles tendinopathy can manifest as scattered punctate hypointensity in areas of high signal intensity. The two radiologists showed a good consistency in measuring the T2*M, T2*S, T2*L and Fraction values of the achilles tendon, and the ICC values ââwere 0.96, 0.94, 0.83 and 0.94, respectively. The T2*s values was significantly higher in the whole Achilles tendon, MTJ and MID segment at 48 h post-exercise compared to pre-exercise, and decreased after 1 month of exercise, [0.49 (0.45, 0.59) vs 0.54 (0.49, 0.59) vs 0.53 (0.49, 0.57), 0.48 (0.44, 0.54) vs 0.53 (0.47, 0.58) vs 0.50 (0.46, 0.57), 0.48 (0.43, 0.58) vs 0.54 (0.47, 0.59) vs 0.52 (0.46, 0.57); respectively, all P<0.05]. The changes in T2*M, T2*L and Fraction values are not statistically significant (all P>0.05). In different running gestures, the ΔT2*S of achilles tendon who using the postures of front-middle feet is higher than that using the postures of back feet (0.03(-0.05, 0.07) vs -0.03(-0.17, 0.11), P=0.001). Conclusion: The Bi-component analysis of UTE-T2* technology is superior to single component analysis in monitoring the dynamic changes of achilles tendon before and after exercise, and T2*S is a more sensitive sequence to evaluate the subtle changes in the chemical composition of achilles tendon.