CT Procedures

When a patient is referred for a CT for suspected urolithiasis, the radiographers must follow corresponding standardized procedure to optimize the radiation dose.

Dose reduction is obtained differently according to the CT device. With definition scanner devices, the reduced dose is obtained through a reduction of the tube current-time product expressed in mAs. Compared with a standard abdominal CT protocol, the reduced dose protocol delivers 33% lower tube current-time product, and its name includes a specific low-dose label (‘LD’). The two remaining CT devices are equipped with a Tin (Sn) filter, enabling a dose reduction of more than 65% compared with the standard abdominal CT protocol. The Tin filter protocol name includes a specific label named ‘Sn’.

Patients and Audit

CT examinations were selected through a DACS (Data Acquisition Computerized System) program containing anonymized dose reports and text of CT reports issued by radiologists. This program provided access to examinations dealing with stone diseases through several key words (‘renal colic’, ‘ureteric stone’, ‘migration’). These examinations were transferred to a dedicated server and data analysis was performed on this server, with complete anonymization of patient’s data except for their age and gender.

Information regarding CT examinations was collected retrospectively and included: patient’s age, examination indication (lithiasis control, renal colic, pain in the right or left iliac fossa), procedure choice, acquisition length, radiation dose descriptors and the radiologist’s name, which was also anonymized.

A first audit period took place from February to April 2021. During May 2021, the radiologist responsible for quality control and radioprotection in the department reminded the stakeholders of the urolithiasis procedure. A second audit analyzed data from June to August 2021.

Image Analysis

Compliance with the procedure was evaluated according to two criteria: the selection of the appropriate acquisition protocol and the reduction of acquisition height with exclusion of the top of the liver. For each exam, we collected the DLP and CTDI_volume.

Statistical Analysis

We calculated the proportion of examinations performed according to the urolithiasis procedure amongst all the urolithiasis CT scans performed during both audit periods and per radiologist. The proportion comparisons were calculated using the χ² Pearson test.

We performed a Shapiro-Wilk test to testify the normality of our quantitative variables.

We compared medians using the Mann-Whitney test. Statistical significance was set for p values < 0.05.

Urolithiasis Procedure

Table 1 summarizes the percentages and proportions of urolithiasis examinations conducted in compliance with the procedure for the three hospitals for both periods.

We reached statistical significance for hospital A and B (p < 0.001 and p = 0.013 respectively) but not for hospital C (p = 0.405).

Figure 1 illustrates 3 examples of CT-scans in the coronal plane from patients with similar diameters and different acquisition lengths and/or protocol selections.

Radiologist Dependency

Table 2 summarizes the percentages and proportions of urolithiasis examinations conducted in compliance with the procedure for each radiologist.

Radiation Dose

The median DLP was 162.6 mGy.cm (p25: 77 | p75: 256.2) for the first audit and 95.2 mGy.cm (p25: 58.65 | p75: 216) for the second audit (p < 0.001).

Audit Interval

Compared with Oliveri’s one-year control, which had reached a compliance of 85%, our first audit conducted five years later shows a drop to 55.65% of the compliance to the urolithiasis procedure and our second audit reached 69.42%.

These observations deserve further discussion.

First, regarding the impact of surveying compliance, our findings are consistent with those reported by Oliveri et al. showing an increased compliance after informing the staff [10]. However, they also demonstrate that the compliance which proved to be stable at one year, decreased thereafter.

Indeed, the percentage of compliance with the urolithiasis procedure had reached 85% by the end of the study, but our first audit, which took place in the same radiology department five years later, reached only 55.65%. This suggests that the effect of one quality control process on compliance decreases over time. To ensure a sustainable improvement with compliance, these audits should be repeated regularly, as required by the Federal Agency for Nuclear Control (FANC) since 2020 [12], or even better, monitored continuously using DACS or similar applications. The optimal delay for repeating the audit is not yet known and would probably depend on several human and organizational factors specific to each department.

Second, even if compliance increased, it does not reach perfection. Indeed, our audit showed an increase in compliance of 15% and reduction of the median DLP by 41.45% (p < 0.001). The reduction achieved in DLP shows that, even if it seems low, it has a real impact on the radiation dose delivered to the patient. However, the DLP can be influenced by the patient diameter, a value which we unfortunately did not have at our disposal.

As illustrated in Figure 1, the dose reduction can be very important, as high as 83%, which should largely motivate stakeholders to use low-dose protocols.

However, even if not perfect, compared with data from the literature, the use of reduced-dose CT in urolithiasis in our department is much higher than reported elsewhere [13]. Indeed, Weisenthal et al. reported that the use of this protocol is infrequent and as low as 5.6%. The corresponding DLP was 689 mGy.cm, seven times higher than in our study.

Thirdly, we found that radiologists showed varying degrees of compliance ranging from 37.5% to 100% for the first audit and from 0% to 88.89% for the second audit.

Other studies have already shown radiologist’s dependence for adhesion to guidelines and found factors associated with a higher rate of compliance, such as practice in a teaching hospital setting or fewer than five years of experience [14]. Another reason that could explain such differences could be a fear of inadequate image quality, even though studies have shown the same diagnostic performance with low-dose CT [6, 7, 8].

Fourthly, compliance with the first criterion (low-dose protocol) is higher than compliance with the second (acquisition length). This could be because low-dose protocol selection requires less interaction on the scanner console than lowering the acquisition length. Some radiologists may also be reluctant to lower the acquisition length by fear of missing an alternative diagnosis, despite the evidence by Brassard et al. that alternative diagnosis are detectable on the restricted abdominal coverage [15].

Lastly, we reached statistical significance for hospital A and B but not for hospital C. This could be because hospital C conducted less scans for urolithiasis during the same period. This has been found to be a significant factor in other articles, such as one by Eisenberg et al., who showed that adhesion to Fleischner guidelines in chest imaging was dependent on the number of CT scans for small nodule detection in each center [14].

One limitation of our study is that we only conducted a single-component intervention. A study by Smith-Bindman showed that detailed feedback on CT radiation dose along with actionable dose-lowering suggestions reduced doses more significantly than an audit feedback alone [16]. We could train radiographers continuously by setting this topic on a regular basis to be discussed in meeting with the staff.

Another limitation is that we were unable to investigate the potential impact of the time of day that the examinations were performed. Is there an impact of night duty or of the number of radiographers present in the department? More data will be available soon, thanks to new functionalities of the DACS.