Clinical Studies of Real‐Time Monitoring of Lithotripter Performance Using Passive Acoustic Sensors Available to Purchase

This paper describes the development and clinical testing of a passive device which monitors the passive acoustic emissions generated within the patient’s body during Extracorporeal Shock Wave Lithotripsy (ESWL). Designed and clinically tested so that it can be operated by a nurse, the device analyses the echoes generated in the body in response to each ESWL shock, and so gives real time shock‐by‐shock feedback on whether the stone was at the focus of the lithotripter, and if so whether the previous shock contributed to stone fragmentation when that shock reached the focus. A shock is defined as being ‘effective’ if these two conditions are satisfied. Not only can the device provide real‐time feedback to the operator, but the trends in shock ‘effectiveness’ can inform treatment. In particular, at any time during the treatment (once a statistically significant number of shocks have been delivered), the percentage of shocks which were ‘effective’ provides a treatment score $TS(t)$ which reflects the effectiveness of the treatment up to that point. The $TS(t)$ figure is automatically delivered by the device without user intervention. Two clinical studies of the device were conducted, the ethics guidelines permitting only use of the value of $TS(t)$ obtained at the end of treatment (this value is termed the treatment score $TS0$⁠). The acoustically‐derived treatment score was compared with the treatment score $CTS2$ given by the consultant urologist at the three‐week patient’s follow‐up appointment. In the first clinical study (phase 1), records could be compared for 30 out of the 118 patients originally recruited, and the results of phase 1 were used to refine the parameter values (the ‘rules’) with which the acoustic device provides its treatment score. These rules were tested in phase 2, for which records were compared for 49 of the 85 patients recruited. Considering just the phase 2 results (since the phase 1 data were used to draw up the ‘rules’ under which phase 2 operated), comparison of the opinion of the urologist at follow‐up with the acoustically derived judgment identified a good correlation $(kappa = 0.94),$ the device demonstrating a sensitivity of 91.7% (in that it correctly predicted 11 of the 12 treatments which the urologist stated had been ‘successful’ at the 3‐week follow‐up), and a specificity of 100% (in that it correctly predicted all of the 37 treatments which the urologist stated had been ‘unsuccessful’ at the 3‐week follow‐up). The ‘gold standard’ opinion of the urologist $(CTS2)$ correlated poorly $(kappa = 0.38)$ with the end‐of‐treatment opinion of the radiographer $(CTS1).$ This is due to the limited resolution of the lithotripter X‐Ray fluoroscopy system. If the results of phase 1 and phase 2 are pooled to form a dataset against which retrospectively to test the rules drawn up in phase 1, when compared with the gold standard $CTS2,$ over the two clinical trials (79 patients) the device‐derived scored $(TS0)$ correctly predicted the clinical effectiveness of the treatment for 78 for the 79 patients (the error occurred on a difficult patient with a high body mass index). In comparison, using the currently available technology the in‐theatre clinician (the radiographer) provided a treatment score $CTS1$ which correctly predicted the outcome of only 61 of the 79 therapies. In particular the passive acoustic device correctly predicted 18 of the 19 treatments that were successful (i.e. 94.7 sensitivity), whilst the current technology enabled the in‐theatre radiographer to predict only 7 of the 19 successful treatments (i.e. 36.8 sensitivity). The real‐time capabilities of the device were used in a preliminary examination of the effect of ventilation.