Standardizing CT-Guided Biopsy Procedures: Patient Dose and Image Noise

Transcription

1 Vascular and Interventional Radiology Original Research Tam et al. Standardization of Image Acquisition Protocol for CT-Guided Biopsy Vascular and Interventional Radiology Original Research FOCUS ON: Alda L. Tam 1 Joe E. Ensor 2 Rodrick C. Zvavanjanja 1,3 Asher Philip 3 Mohamed E. Abdelsalam 1 Steven Y. Huang 1 Sanjay Gupta 1 A. Kyle Jones 4 Tam AL, Ensor JE, Zvavanjanja RC, et al. Keywords: biopsy, CT, CT-guided biopsy, standardization DOI: /AJR Received June 18, 2014; accepted after revision February 4, The University of Texas MD Anderson Cancer Center is supported in part by Cancer Center Support Grant P30CA from the National Cancer Institute, National Institutes of Health. Based on a presentation at the Radiological Society of North America 2012 annual meeting, Chicago, IL. 1 Department of Interventional Radiology, The University of Texas MD Anderson Cancer Center, Houston, TX. 2 Current address: Houston Methodist Cancer Center, Houston Methodist Research Institute, Houston, TX. 3 Department of Diagnostic and Interventional Imaging, The University of Texas Medical School at Houston, Houston, TX. 4 Department of Imaging Physics, The University of Texas MD Anderson Cancer Center, 1400 Pressler, Unit 1472, Houston, TX Address correspondence to A. K. Jones (kyle.jones@mdanderson.org). This article is available for credit. WEB This is a web exclusive article. Supplemental Data Available online at AJR 2015; 205:W390 W X/15/2054 W390 JOURNAL American Roentgen Ray Society CLUB JOURNAL CLUB: Standardizing CT-Guided Biopsy Procedures: Patient Dose and Image Noise OBJECTIVE. The objective of this study was to standardize our image acquisition protocol for CT-guided biopsy procedures. MATERIALS AND METHODS. The records of consecutive patients who underwent CT-guided biopsy 3 months before (n = 598 biopsies) and 3 months after (n = 540 biopsies) standardization of our image acquisition protocol were retrospectively reviewed. CT technical parameters were individualized on the basis of the sum of the anteroposterior and transverse dimensions of the patient. Information on patient demographic characteristics, biopsy site, complications associated with the procedure, and diagnostic yield was collected. The radiation dose metrics that were evaluated included the volume CT dose index, dose-length product, and size-specific dose estimate. Image noise was quantified using the SD of the CT number measured in subcutaneous fat. Fisher exact test and one-way ANOVA were used to evaluate statistical significance. RESULTS. The mean dose-length product decreased by 72.3% (from to mgy cm; p < ), and statistically significant decreases in dose-length product were observed when data were stratified according to biopsy site (i.e., lung, solid organ, lymph node, or bone; for all sites, p < ). The mean size-specific dose estimate decreased by 58.9% (from 125 to 51.4 mgy), which was statistically significant (p < 0.001). Image noise increased during the study period, but this increase was not statistically significantly different among the four biopsy sites (p = 0.46). CONCLUSION. Standardization of the image acquisition protocol used in CT-guided biopsy procedures significantly reduced patient radiation dose and decreased variability in image noise. T he radiation dose associated with medical imaging is an issue affecting patient safety and a topic of growing public interest. The National Council on Radiation Protection and Measurements reported that the per capita radiation dose resulting from medical imaging has increased by a factor of almost 6 in the United States since the early 1980s and estimated that CT accounts for approximately 25% of the annual cumulative radiation dose to the population [1]. Although CTguided interventions represent only a small percentage of the total number of CT scans performed each year, the number of CTguided interventions is increasing, and substantial radiation doses can be delivered during these examinations [2, 3]. Investigators are beginning to explore methods to reduce radiation dose during CT-guided interventions, with some studies focusing on the use of CT fluoroscopy [4 7] and others assessing the use of conventional CT techniques [3, 8 10]. Regardless of the method of CT guidance used during the intervention, these dose reduction techniques have relied on either arbitrarily decreasing the tube current exposure time product or using a weight-based algorithm [3, 8 10]. For diagnostic CT examinations, technical parameters such as kilovoltage and tube current time product settings are carefully tailored to individual patients on the basis of their body size [3, 4], but this approach is not applied uniformly in CT-guided interventional procedures. An effective strategy for radiation dose reduction in CT-guided interventions requires that the volume imaged be restricted to the target of the intervention and that CT parameters be adapted to patient dimensions to reduce both overall radiation dose and the variability in dose between in- W390 AJR:205, October 2015

2 Standardization of Image Acquisition Protocol for CT-Guided Biopsy dividual examinations of the same type [11]. The practice at our institution was to use the default CT parameters for CT-guided biopsy procedures, including biopsy phase techniques of 120 kv and 50 mas and preprocedural planning scan (PPS) techniques of 120 kv and 150 mas for the chest and lung, 120 kv and 200 mas for the abdomen and pelvis, and 120 kv and 300 mas for the spine. Recognizing that the image quality required for an image-guided intervention does not have to be equivalent to that for a diagnostic imaging study and that technical factors should be adapted to individual patients during CT-guided interventional procedures, we sought to standardize our image acquisition protocol for CT-guided biopsy procedures. We developed a method with which to tailor the CT parameters to the dimensions of the individual patient and the biopsy region, and we integrated this method into the workflow through the use of a protocol posted at each CT scanner. The purpose of this study was to evaluate the effect of standardization of our image acquisition protocol for CT-guided biopsies on image noise and radiation dose metrics, including volume CT dose index (CTDI vol ), dose-length product (DLP), and size-specific dose estimate (SSDE) [12]. Materials and Methods Study Design The quality improvement assessment board at our institution (The University of Texas MD Anderson Cancer Center, Houston, TX) approved this retrospective data review study. In September 2011, the interventional radiology department standardized the image acquisition protocol for CT-guided biopsy procedures. We retrospectively compared data collected from patients who underwent CT-guided biopsy during a 3-month period after standardization (defined as the study period, which lasted from November 14, 2011, through February 14, 2012) with data collected from patients who underwent CT-guided biopsy during the 3 months before standardization (defined as the prestudy period, which lasted from June 1, 2011, through August, 31, 2011). Consecutive adult patients who underwent CT-guided biopsies were included in the study. Biopsies of the head and neck or the extremities and biopsies performed on pediatric patients were excluded from the study (n = 56). Therefore, a total of 1138 biopsy procedures were included in this analysis: 598 biopsies were performed for 577 patients (307 men and 270 women; mean age, 60.7 years) during the prestudy period, and 540 biopsies were performed for 515 patients (260 men and 255 women; mean age, 60.3 years) during the study period. Standardization of CT Image Acquisition The CT-guided biopsy procedure was divided into a series of six steps, and a protocol listing these steps was placed at each CT scanner (Fig. 1). The protocol specified the length of the PPS and provided instructions and charts for selecting CT parameters. The CT parameters were selected on the basis of the sum of anteroposterior and transverse dimensions of the patient, as measured on a previously obtained diagnostic cross-sectional image depicting the lesion to be biopsied. Aside from these two changes, the procedural steps for performing CT-guided biopsy were identical to the steps we had previously used. Patients were positioned on the CT gantry, and a radiopaque marker was placed on the skin overlying the expected puncture site. Kilovoltage and tube current time product settings were selected from the aforementioned CT parameter charts and were used for acquisition of the PPS, which was obtained using helical scanning mode, and for acquisition of all subsequent images during the biopsy procedure. The needle path for the biopsy procedure was selected, and the procedure was performed using sterile technique. To evaluate the introduction, progression, and placement of the biopsy needle, intermittent CT scans were obtained using biopsy mode (in which images were acquired using axial scanning mode) during the procedure. Biopsy procedures were performed by one of 10 board-certified interventional radiologists, all of Fig. 1 Copy of unique chart placed at each CT scanner to show standardized CT image acquisition protocol used during biopsy procedures. Technologists were instructed to limit preprocedural planning scan length to 75 mm. Lung chart was used for all biopsies of chest, including those involving lung, lymph node, and bone. AP = anteroposterior, Trans = transverse, EMR = electronic medical record. AJR:205, October 2015 W391

3 Tam et al. whom had at least 5 years of biopsy experience. The CT-guided biopsies were performed using three CT scanners: a 20-MDCT (Somatom Sensation Open 20, Siemens Healthcare), a 16-MDCT (Somatom Sensation 16, Siemens Healthcare), and a 64-MDCT (Somatom Sensation 64, Siemens Healthcare). Figure 2 illustrates the CT technologist workflow for the standardized CT image acquisition protocol. Adaptation of CT Technical Factors on the Basis of Patient Size A simple model of attenuation and noise was used to create CT technical factor charts for adjusting the kilovoltage and tube current time product settings on the basis of the measured dimensions of the patient, with the goal of maintaining constant image noise. The sum of the anteroposterior and transverse dimensions of the patient, as measured on a cross-sectional image, was used to represent patient size. This method was chosen because it required the technologist simply to add two numbers rather than calculate a quantity such as the effective diameter, which would require the use of a calculator to perform a square root operation. The effective diameter of an object is the geometric mean of the anteroposterior and transverse dimensions of the object, calculated as follows: Effective Anteroposterior (cm) diameter (cm) = Transverse (cm). (1) Using the sum of the anteroposterior and transverse dimensions is equivalent to using the effective diameter, because the two quantities are linearly related [12]. The scaling factor used to create chart templates was as follows: μx e 2, (2) where x was the sum of the anteroposterior and transverse dimensions of the patient and where μ represented the attenuation of the body region imaged. Scaling factors were calculated for anteroposterior and transverse dimensions ranging from 16 to 90 cm with the use of this model, and factors were normalized to an anteroposterior and transverse dimension of 64 cm. The soft-tissue composition published by the International Commission on Radiation Units and Measurements [13] was used to model soft tissue. Scaling factors for lung biopsy were created using a mix of 50% soft tissue with a density of 1.06 g/cm 3 and 50% soft tissue with a density of 0.3 g/cm 3 (to mimic inflated lung tissue), for calculation of the body region attenuation value (see Table S1; Tables S1 S3 and Fig. S4, supplemental material, can be viewed in the AJR electronic supplement to this article, available at These fractions were based on measurements made from a series of CT images of actual patients. Scaling factors for biopsy of the abdomen or pelvis were created using 100% soft tissue with a density of 1.06 g/cm 3 for calculation of the body region attenuation value (Table S2). These scaling factors were also used to create a chart with a higher baseline tube current time product setting for solid-organ biopsy. Unique charts were created for each of our three CT scanners, with the goal of maintaining the same scanner output by accounting for differences in x-ray output as a function of kilovoltage (using the weighted CT dose index [CTDI w ] measured in the 32-cm CTDI phantom) and for differences in generator power (which limits the maximum tube current time product at each kilovoltage setting). A further refinement to our CT parameter charts was made after the study period: reduced kilovoltage settings were added for small patient dimensions (Fig. S1). Fixed tube current time product settings were chosen over automatic tube current modulation because of the limited craniocaudal length of interventional radiology scans [14] and the limited customization options available for automatic tube current modulation on our CT scanners. Acceptable baseline CT parameters were determined by surveying all interventional radiologists. The baseline parameters for a standard patient with an anteroposterior and transverse dimension of 64 cm were 120 kvp and 50 mas for chest biopsy and 120 kvp and 100 mas for biopsy of the abdomen or pelvis performed using the 64-MDCT scanner. Baseline parameters and parameter charts were adjusted for each of the three CT scanners, to account for differences in scanner radiation output. Data Collection Data collected from the electronic medical record included age, sex, body mass index (BMI; weight in kilograms divided by the square of height in meters), biopsy site, complications associated with the biopsy procedure, pathologic findings, and the interventional radiologist who performed the procedure. Biopsies were stratified by site and biopsy region analyzed: lung, lymph node (chest and abdomen or pelvis), solid organ, and bone (chest and abdomen or pelvis). Solid organs included the liver, spleen, kidney, and adrenal gland. Data were also collected from the radiology PACS and included the times that the first and last images were obtained, PPS length, anteroposterior and transverse dimensions, number of biopsy images, biopsy image thickness, reconstruction kernel used for biopsy images, total DLP used for the procedure, and total CTDI vol for the images obtained during biopsy needle placement. To evaluate the effect of the standardized protocol on image noise, we measured the SD of CT numbers in the subcutaneous fat in three consecutive biopsy images: one image documenting the biopsy needle within the lesion plus one image superior to and one image inferior to the center image. Measurements were performed so as to avoid metal artifacts from the needle. The mean of the three measurements was calculated and normalized to a single reconstruction kernel (B31s) and image thickness (4.8 mm). Normalization factors for reconstruction kernels and image thickness were calculated using measurements of image noise in a uniform water phantom (Table S3). To evaluate the effect of the standardized protocol on radiation dose, we analyzed the following metrics: total DLP, fraction of DLP attributed to needle placement and biopsy sampling, total SSDE, and SSDE normalized per biopsy scan. The total SSDE was equal to the sum of the SSDEs for scans acquired during all phases of the procedure, and the SSDE per biopsy scan was equal to the sum of the SSDEs for all scans obtained during the biopsy phase divided by the number of scans acquired during the biopsy phase. The DLP is most closely related to the total energyimparted to the patient and includes contributions from both scanner output (i.e., CTDI vol ) and the volume of tissue irradiated (i.e., scan length). The SSDE is a dose metric that takes into account the size of the patient [12]: the SSDE scales the reported CTDI vol, which is related to radiation output from the CT scanner, to account for differences in absorbed dose to critical organs as a function of patient size. Deviations from the standardized protocol were recorded and analyzed. Statistical Analysis Fisher exact test was used to evaluate categoric data. Differences in means were tested using oneway ANOVA, and differences in variation were assessed using a folded F-test. Shapiro-Wilk test was used to assess the normality of the data objectively, and these findings were confirmed visually by inspecting the normal probability quantile-quantile plots. In instances where distributions were nonnormal, a log transformation was applied to the data and normality was reassessed. Statistical significance was defined as p < All analyses were conducted using SAS software (version 9.3, SAS Institute). Ordinary least-squares regression analysis was used to evaluate the relationship between either SSDE per biopsy scan or normalized image noise and the effective diameter of the patient within each period. Model differences between the two periods were assessed on the basis of results of Satterthwaite t tests, accounting for the difference in residual variation between the two periods. Results Tissue was sampled successfully in all biopsy procedures, and there were no differences be- W392 AJR:205, October 2015

4 Standardization of Image Acquisition Protocol for CT-Guided Biopsy TABLE 1: Effect of Standardizing the Image Acquisition Protocol for CT-Guided Biopsy Procedures on Dose-Length Product (DLP) and Size-Specific Dose Estimate (SSDE) Per Biopsy Scan Biopsy Site, Region, and Study Period Lymph node No. of Biopsies Performed Geometric Mean DLP (mgy cm) Change in Geometric Mean From Prestudy to Study Period (%) p Geometric Mean SSDE Per Biopsy Scan (mgy) Change in Geometric Mean From Prestudy to Study Period (%) Chest < < Prestudy Study Abdomen or pelvis < < Prestudy Study Bone Chest < < Prestudy Study Abdomen or pelvis < < Prestudy Study Lung, chest < < Prestudy Study Solid organ, a abdomen or pelvis < < Prestudy Study Total < < Prestudy Study a Includes liver, kidney, adrenal gland, and spleen biopsies. Study Period Prestudy Period Dose-Length Product (mgy cm) A Study Period Prestudy Period Size-Specific Dose Estimate per Rotation (mgy cm) Fig. 2 Boxplots with whiskers denoting minimum and maximum values and boxes denoting interquartile ranges. A, Boxplot of dose-length product data on log scale. Data show statistically significant decrease both in geometric mean dose-length product (diamonds) between two periods (p < ) and in SD (width of gray box) between two periods (p < ). B, Boxplot of data on size-specific dose estimate per rotation of the tube on log scale. Data show statistically significant decrease both in geometric mean (diamonds) between two periods (p < ) and in SD (width of gray box) between two periods (p < ). p B AJR:205, October 2015 W393

5 Tam et al. tween the prestudy and study periods in terms of the complication rate (13% [78/598 biopsy procedures] vs 10.6% [57/540 biopsy procedures]; p = 0.20) or the diagnostic yield (94.3% [546/579 biopsy procedures] vs 95.8% [500/522 biopsy procedures]; p = 0.27). There was no statistically significant difference in the body size of patients between the prestudy and study periods: the mean BMI was 28.1 (range, ) versus 28.2 (range, ) (p = 0.94), the mean anteroposterior dimension was versus mm (p = 0.83), and the mean transverse dimension was versus mm (p = 0.70). Implementation of the standardized CT protocol did not adversely affect work flow: the time required to complete the biopsy procedure was statistically significantly less in the study period than in the prestudy period (31.6 vs 34.8 minutes; p = 0.038). Effect on Radiation Dose The total DLP decreased by 72.3% (p < ) between the two periods, and statistically significant reductions in the DLP were also observed when the biopsy procedures were stratified by biopsy site (Table 1). The interbiopsy variation in total DLP decreased between periods (Fig. 2), indicating that use of the standardized CT protocol reduced variability in the performance of CT-guided biopsies. The decrease in the mean DLP was driven by a reduction in both the mean length of the PPS (geometric mean, 187 mm during the prestudy period vs 110 mm during the study period; p < ) and the mean SSDE per biopsy scan (geometric mean, 10.5 mgy during the prestudy period vs 4.58 mgy during the study period; p < ). This finding indicates that, on average, a smaller volume of the patient was imaged using less radiation. Statistically significant reductions in the SSDE per biopsy scan were also observed when biopsy procedures were stratified by biopsy site (Table 1), and a statistically significant decrease in total SSDE was noted (geometric mean, 125 mgy during the prestudy period vs 51.4 mgy during the study period; p < ). On average, in the prestudy period, 42% of the total DLP for a procedure was attributed to the biopsy phase. This fraction increased to 59% during the study period (p < ), occurring as a result of the decrease in DLP associated with the reduced PPS length. Effect on Image Noise A total of 1053 biopsy procedures were evaluated as part of the image noise analysis: 596 biopsies in the prestudy period and 457 biopsies in the study period. We excluded from our analysis 83 biopsies that were performed during the study period, because the technical parameters used for these procedures deviated from the standardized protocol. In these cases, a higher- or lower-than-indicated tube current time product setting or kilovoltage setting or both were used during the biopsy procedure, as a result of either a physician request or a mistake made by the technologist. The SD of image noise decreased for lymph node biopsies of the chest and of the abdomen or pelvis and for bone biopsies of the abdomen or pelvis; the decrease was statistically significant for lymph node and bone biopsies of the abdomen or pelvis (p = and p = , respectively), indicating increased standardization of image noise (Table 2). Overall, image noise increased statistically significantly between the prestudy and study periods (from 13.3 to 19.6 HU; p < ), as a consequence of the overall reduction in the SSDE per biopsy scan (Fig. 3). Regression analysis of the image noise data showed a statistically significant correlation with patient effective diameter in the prestudy period (regression slope, 0.724; p < ) and a decreased correlation with patient effective diameter in the study period (regression slope, 0.632; p = 0.45) for biopsies of the chest; however, the decrease was not statistically significant. Regression analysis of the image noise data for biopsies of the abdomen or pelvis revealed a statistically significant correlation with patient effective diameter in the prestudy period (regression slope, 0.538; p < ) and a statistically significantly decreased correlation with patient effective diameter in the study period (regression slope, 0.108; p = 0.004). Deviation Analysis Complying with all six elements of the protocol was challenging for both the technologists and the interventional radiologists. The rate of deviation from the standardized protocol during the study period was 72.6% (392 of 540 biopsy procedures). Table 3 details the reasons for deviation. Technologists accounted for 78% of deviations (304/392), physicians for 12% (46/392), and multiple causes TABLE 2: Effect of Standardizing the Image Acquisition Protocol for CT-Guided Biopsy Procedures on Normalized Image Noise Biopsy Site, Region, and Study Period No. of Biopsies Performed Normalized Image Noise (HU), Mean ± SD Lymph node Chest 0.16 Prestudy ± 8.12 Study ± 5.70 Abdomen or pelvis Prestudy ± 7.77 Study ± 6.04 Bone Chest 0.14 Prestudy ± 6.31 Study ± 8.70 Abdomen or pelvis Prestudy ± 10.1 Study ± 6.04 Lung, chest < b Prestudy ± 4.87 Study ± 7.40 Solid organ c, abdomen or pelvis 0.95 Prestudy ± 4.83 Study ± 4.85 a Folded F-test. b Although this p value is statistically significant, the SD was higher in the study period. c Includes liver, kidney, adrenal gland, and spleen biopsies. p a W394 AJR:205, October 2015

6 Standardization of Image Acquisition Protocol for CT-Guided Biopsy (e.g., technologist and physician) for 10% (39/392). Less than 1% of deviations (3/392) resulted from complications. Common compliance problems included PPS lengths longer than 75 mm, failure to use the kilovoltage and tube current time product settings indicated for the patient on the basis of his or her size, and requests from physicians for higher tube current time product settings, which occurred in 8.1% of biopsy cases (44/540) and, in particular, during biopsies of solid organs. Discussion All specialists, including interventional radiologists, have a responsibility to ensure that an appropriate radiation dose is used for medical imaging. This can be achieved by scaling CT parameters to patient body size and to the type of study being performed, for example, diagnostic imaging versus imaging for guidance during a procedure [3, 4, 15 17]. Before the introduction of this standardized protocol at our institution, the CT parameters used during biopsy procedures either were the default parameters installed on the scanner or were selected based on a technologist s visual assessment of patient size, and the length of the PPS was not specified. The mean DLP (700 mgy cm) for biopsy procedures performed during the prestudy period was comparable to that for typical diagnostic CT [18, 19]; however, this DLP is not unusual for CT-guided biopsy procedures [2, 20]. At our institution, an estimated 2800 CTguided biopsy procedures are performed annually, and standardization of the procedure, which resulted in a mean 72.3% decrease in total DLP and a mean 61.8% decrease in total SSDE, is expected to improve patient safety. The results of our regression analysis indicated that normalized image noise was relatively constant, regardless of patient effective diameter, for biopsies of the abdomen or pelvis but not for biopsies of the chest. This finding was also reflected in the SD of image noise (Table 2). Two potential causes of this discrepancy were identified. First, the CT parameter charts lacked reduced kilovoltage settings that would have been used for patients with dimensions smaller than 32 cm. Second, attenuation in the chest, and therefore image noise, are affected by lung density, which is directly related to lung inflation. Breathing cycles and the depth of respiration vary from patient to patient, and they are more likely to be erratic during a biopsy procedure. The SD of image noise for solid-organ biopsies was constant between the prestudy and study periods, potentially a consequence of in- Fig. 3 Biopsy images for three patients who underwent CT-guided biopsy in both prestudy and study periods. Study period images show increase in image noise, but biopsy needle, target lesion, and adjacent structures are clearly visualized. A, 58-year-old woman with non small cell lung cancer of left upper lobe, body mass index (BMI [weight in kilograms divided by the square of height in meters]) of 19.9, and effective diameter of 23.8 cm. Dose-length product (DLP) was 489 mgy cm in prestudy period versus 69 mgy cm in study period, and size-specific dose estimate (SSDE) per biopsy scan decreased from 11.6 to 2.92 mgy in same periods, respectively. B, 63-year-old woman with history of lymphoma and mesenteric mass, BMI of 24.6, and effective diameter of 33.7 cm. DLP was 758 mgy cm in prestudy period versus 104 mgy cm in study period, and SSDE per biopsy scan decreased from 11.2 to 3.75 mgy in same periods, respectively. C, 59-year-old woman with history of lymphoma and bilateral retroperitoneal lymphadenopathy, BMI of 34.9, and effective diameter of 32.8 cm. DLP was 636 mgy cm in prestudy period for biopsy of 3-cm right retroperitoneal lymph node versus 173 mgy cm in study period for biopsy of 1-cm left paraaortic lymph node; SSDE per biopsy scan decreased from 14.3 mgy to 4.25 Gy. A B C AJR:205, October 2015 W395

7 Tam et al. TABLE 3: Analysis of Deviations From the Standardized Protocol During the Study Period Reason Technologist Physician Technologist and Physician A B C A and B A and C B and C A and B and C Total Note Data are number of deviations. A = preprocedural planning scan (PPS) too long, B = repeated PPS, C = wrong technique. cluding solid-organ biopsies from the study period that were not recorded as deviations (because of the use of higher tube current time product) in the image noise analysis. Our approach to standardization of CT image acquisition was unique for several reasons. By tailoring the CT parameters to each patient s dimensions in the plane of the lesion to be biopsied and then adapting them to the specific characteristics of each CT scanner, our method scales CT parameters to the patient more precisely than do strategies that rely on weight-based methods or arbitrary reduction of the tube current time product. The increased correlation between SSDE per biopsy scan and patient effective diameter and the decreased correlation between normalized image noise and patient effective diameter indicated that technical factors and radiation dose were better matched to patient size in the study period than in the prestudy period. Image noise was also standardized in this study, as is shown by the decrease in the SD of image noise between periods. Furthermore, the scaled CT parameters were used for both the PPS and the images acquired during the biopsy phase of the procedure, which is different from studies that used a higher tube current time product for the PPS and a reduced tube current time product for the biopsy phase of the procedure [4, 5, 8]. Last, the PPS length was restricted to 75 mm in the standardized protocol. Long PPSs contribute to increased radiation dose, yet previous studies did not explicitly restrict the PPS A length; instead, they did not address the PPS [2 4, 7, 8, 10], limited the PPS to the area of interest [5, 6], or had the radiologist determine the area to be scanned [9]. The effect of not limiting the PPS is made evident by the fact that, in these previous studies, the PPS was reported to contribute 66.4% 89% of the total radiation dose resulting from CT-guided biopsy procedures [2, 4, 20, 21]. In our study, the percentage of DLP attributed to the PPS decreased from 58% in the prestudy period to 41% during the study period. Technologists attributed the difficulty in complying with the PPS length limit to wanting to ensure that the target lesion was captured during the PPS. In response to the study results and their feedback, we configured our CT scanners to set the PPS length to 75 mm by default, and we reviewed anatomic landmarks with the technologists to increase their comfort level in determining the area to be scanned. There was an initial adjustment period during which the technologists became familiar with how to measure the dimensions of a patient, thus accounting for some of the errors in selecting the correct techniques. Many of the deviations in technique attributed to physicians occurred during biopsies of solid organs. In response to these results, an additional CT parameter chart for solid-organ biopsies, with a baseline of 150 mas, was added to the standardized protocol (Fig. S1). This study was not without some limitations and challenges. Despite the large number of bi- Fig. 4 Illustration of CT technologist workflow when using standardized CT image acquisition protocol. A, Diagnostic imaging study showing lesion to be biopsied (green arrowhead) is accessed in electronic medical record by technologist. Scout line mode is used to determine location of lesion on diagnostic topogram and to determine anatomic landmarks that are to be used to center preprocedural planning scan (PPS) during biopsy. B, Diagnostic image showing lesion (green arrowhead) to be biopsied is used to measure and sum anteroposterior and transverse dimensions of patient. In this case, sum of anteroposterior and transverse dimensions is 22 cm + 40 cm = 62 cm. (Fig. 4 continues on next page) B W396 AJR:205, October 2015

8 Standardization of Image Acquisition Protocol for CT-Guided Biopsy Fig. 4 (continued) Illustration of CT technologist workflow when using standardized CT image acquisition protocol. C, CT parameter chart for lung used by technologist to determine parameters for acquisition of image of pleural lesion. For patient with anteroposterior and transverse dimension of 62 cm, CT parameters for this biopsy procedure should be 120 kvp and 44 mas. AP = anteroposterior, Trans = transverse. D, Screen interface seen by technologist during biopsy procedure showing that PPS range and CT technical factors have been set according to standardized CT image acquisition protocol. C opsies included in the analysis, this study was a retrospective evaluation performed at a single institution. Reproducing these results with the same standardized protocol at other institutions would validate and strengthen the conclusions of this study. There was resistance from some interventional radiologists concerning the increase in image noise. During a biopsy procedure, the image quality must be adequate to target the lesion, avoid critical structures, and localize the biopsy needle; it need not be at the same level of quality used for diagnostic CT (Fig. 4). Education and awareness will be necessary to define adequate image quality and the acceptable tradeoffs between reducing radiation dose and maintaining necessary image quality during CT-guided biopsy. Finally, there were shortcomings to our initial attempts at creating CT parameter charts, most notably a lack of reduced kilovoltage settings for patients with small dimensions (e.g., patients with lung anteroposterior and transverse dimensions of < 47 cm, as shown in Fig. 1), settings that would occasionally have been used, if available. Despite these limitations, the implementation of a standardized image acquisition protocol, which can be easily adopted by other practices, significantly decreased patient radiation dose, reduced variability in individual biopsy procedures performed at the same site, and standardized image noise, all without negatively affecting the duration of the procedure, the complication rate, or the diagnostic yield. Our experience highlights the importance of educating technologists and soliciting physician input about acceptable image noise levels for different types of biopsy procedures. Prospectively involving all stakeholders increases the likelihood of such a standardized protocol being adopted and used long term, and it will increase the effect of the protocol on radiation dose and image quality standardization. References 1. National Council on Radiation Protection and Measurements (NCRP). NCRP report no. 60: ionizing radiation exposure of the population of the United States. NCRP website. org/publications/press_releases/160press.html. Published Accessed June 9, Leng S, Christner JA, Carlson SK, et al. Radiation dose levels for interventional CT procedures. AJR 2011; 197:[web]W97 W Dixon RG, Ogden K. Optimizing dose in computed tomographic guided procedures. Tech Vasc Interv Radiol 2010; 13: Teeuwisse WM, Geleijns J, Broerse JJ, Obermann WR, Van Persijn Van Meerten EL. Patient and staff dose during CT guided biopsy, drainage and coagulation. Br J Radiol 2001; 74: Chintapalli KN, Montgomery RS, Hatab M, Katabathina VS, Guiy K. Radiation dose management. Part 1. Minimizing radiation dose in CTguided procedures. AJR 2012; 198:W347 W Paulson EK, Sheafor DH, Enterline DS, McAdams HP, Yoshizumi TT. CT fluoroscopy-guided interventional procedures: techniques and radiation dose to radiologists. Radiology 2001; 220: Yamao Y, Yamakado K, Takaki H, et al. Optimal scan parameters for CT fluoroscopy in lung interventional radiologic procedures: relationship between radiation dose and image quality. Radiology 2010; 255: Lucey BC, Varghese JC, Hochberg A, Blake MA, Soto JA. CT-guided intervention with low radiation dose: feasibility and experience. AJR 2007; 188: Smith JC, Jin DH, Watkins GE, Miller TR, Karst JG, Oyoyo UE. Ultra-low-dose protocol for CT-guided lung biopsies. J Vasc Interv Radiol 2011; 22: Shepherd TM, Hess CP, Chin CT, Gould R, Dillon WP. Reducing patient radiation dose during CTguided procedures: demonstration in spinal injections for pain. AJNR 2011; 32: D AJR:205, October 2015 W397

9 Tam et al. 11. Sarti M, Brehmer WP, Gay SB. Low-dose techniques in CT-guided interventions. RadioGraphics 2012; 32: American Association of Physicists in Medicine (AAPM). AAPM report 204: size-specific dose estimates (SSDE) in pediatric and adult body CT examinations. AAPM website. pubs/reports/rpt_204.pdf. Published Accessed June 9, International Commission on Radiation Units and Measurements (ICRU). Tissue substitutes in radiation dosimetry and measurement (report no. 44). ICRU website. tissue-substitutes-in-radiation-dosimetry-andmeasurement-report-44. Published Accessed June 16, Lamba R. Radiation dose optimization for CTguided interventional procedures in the abdomen and pelvis. J Am Coll Radiol 2014; 11: Brenner DJ, Hall EJ. Computed tomography: an increasing source of radiation exposure. N Engl J Med 2007; 357: International Commission on Radiological Protection. Radiation protection in medicine: ICRP publication 105. Ann ICRP 2007; 37: Valentin J; International Commission on Radiation Protection. Managing patient dose in multidetector computed tomography (MDCT): ICRP publication 102. Ann ICRP 2007; 37: Conference of Radiation Control Program Directors. Publication number NEXT_ _T: NEXT computed tomography Conference of Radiation Control Program Directors website www. crcpd.org/pubs/nexttrifolds/next_ ct_t. pdf. Published Accessed June 9, European Commission. Publication EUR EN: European guidelines on quality criteria for computed tomography. European Commission website. Published Accessed June 9, Kloeckner R, dos Santos DP, Schneider J, Kara L, Dueber C, Pitton MB. Radiation exposure in CT-guided interventions. Eur J Radiol 2013; 82: Tsalafoutas IA, Tsapaki V, Triantopoulou C, Gorantonaki A, Papailiou J. CT-guided interventional procedures without CT fluoroscopy assistance: patient effective dose and absorbed dose considerations. AJR 2007; 188: FOR YOUR INFORMATION This article is available for CME and Self-Assessment (SA-CME) credit that satisfies Part II requirements for maintenance of certification (MOC). To access the examination for this article, follow the prompts. The data supplement accompanying this web exclusive article can be viewed by clicking Supplemental at the top of the article. This article has been selected for AJR Journal Club activity. The accompanying Journal Club study guide can be found on the following page. W398 AJR:205, October 2015

10 Standardization of Image Acquisition Protocol for CT-Guided Biopsy APPENDIX 1: AJR JOURNAL CLUB Study Guide Standardizing CT-Guided Biopsy Procedures: Patient Dose and Image Noise Joseph J. Budovec 1, Margaret Mulligan 1, Alan Mautz 2 1 Medical College of Wisconsin, Milwaukee, WI. 2 The Aroostook Medical Center, Presque Isle, ME. jbudovec@mcw.edu, mmulliga@mcw.edu, amautz@emhs.org* Introduction 1. What is the purpose of the study? 2. How would your formally state the authors hypothesis? How would you state the alternative hypothesis? 3. Is this study timely and relevant? Methods 4. What were the inclusion criteria for the study? What were the exclusion criteria? 5. What were the limitations of this study? Were these limitations adequately addressed? 6. What variables were selected for analysis? 7. What statistical methods were used in the study analysis? Results 8. Were the clinical questions answered? Were the hypotheses resolved? 9. Was the sample size large enough to adequately evaluate the protocol changes? Physics 10. How do the volume CT dose index, dose-length product, and size-specific dose estimates differ? 11. Other than the kilovoltage and tube current exposure time product, what CT parameters can be changed to reduce the radiation dose? Discussion 12. At your institution or practice, what processes (if any) are in place to reduce radiation dose exposure during CT-guided biopsies? 13. How do you define adequate image quality? 14. Does your institution or practice alter CT parameters to improve or alter image quality? If so, which parameters are altered, and will this study affect that practice? 15. The study describes standardizing the CT technique used for biopsy procedures of the chest and of the abdomen, pelvis, or both. Do CT technologists at your institution use standardized techniques for routine CT examinations? How are such techniques determined? 16. Who should assume primary responsibility for the cumulative radiation dose to which patients are exposed? Background Reading 1. Chintapalli KN, Montgomery RS, Hatab M, Katabathina VS, Guiy K. Radiation dose management. Part 1. Minimizing radiation dose in CT-guided procedures. AJR 2012; 198:[web]W347 W Sarti M, Brehmer WP, Gay SB. Low-dose techniques in CT-guided interventions. RadioGraphics 2012; 32: *Please note that the authors of the Study Guide are distinct from those of the companion article. AJR:205, October 2015 W399