Optimizing Colon Identification by Window Setting Modulation on Noncontrast Computed Tomography Prior to Percutaneous Nephrolithotomy

Introduction

Colonic perforation is a rare complication of percutaneous nephrolithotomy (PCNL) occurring in about 0.3%–0.6% of procedures, ¹ and it represents a disastrous event for both patients and surgeons. Colonic injury requires special therapeutic measures, including separation of the colon from the urinary tract, broad spectrum antibiotics, and possible laparotomy for colonic repair and diversion, resulting in prolonged hospitalization and postoperative rehabilitation. ² To avoid this complication, preoperative planning based on assessment of cross-sectional imaging, commonly noncontrast computed tomography (NCCT), is valuable. ³ NCCT provides information not only about renal anatomy and stone characteristics but also about the anatomical relationships of the kidney with the surrounding organs (e.g., colon, liver, spleen, and pleural cavity). ⁴ Nevertheless, the lack of contrast material, collapsed bowel segments, and previous intraabdominal operations may limit proper colonic recognition.

The abdomen NCCT window setting is the most used for pre-PCNL planning, but the contribution of other settings in this scenario has not yet been assessed. Manipulation of window settings permits altering the CT grayscale features for highlighting particular structures. Wide window is used in areas of acute differing attenuation values, such as lungs, where alveoli with gas are proximate to vessels, whereas narrow window is excellent for areas of similar attenuation, such as soft tissue. ⁵ The bowel is a tubular hollowed soft tissue organ containing a significant amount of gas within the lumen. These colonic features led to our rationale that a window setting designed for better distinction of gas, such as lung window, may improve identification of colon. The aim of our study is to investigate which window setting may optimize colonic identification for pre-PCNL planning.

Methods

Study protocol

The study was approved by the Institutional Review Board. The study consisted of 11 consecutive preoperative NCCT scans of patients treated in our institution who underwent PCNL. They were performed on a multidetector CT scanner (Brilliance; Philips Medical Systems, Cleveland, OH, USA) with 64 detector rows. The reconstructed slice thickness was 2.0 with an increment of 1.0 mm. The NCCT protocol consisted of scanning from top of the kidneys through bladder base, preferably in one breath hold, with no oral or IV contrast, with the patient in prone position to facilitate distinction between a stone located at the ureterovesical junction and one located in the bladder. The director of our endourology unit and a senior radiologist (M.S. and G.A.) selected the most representative axial NCCT images for planning the renal puncture axis. Each chosen image was retrieved in both abdomen and lung window settings. These images were presented in a random order to 10 senior faculty members and residents in our university-affiliated Urology Department who were informed that the scans had been performed in patients before PCNL (Fig. 1). Following a warm-up session on unrelated NCCT scans, they were asked to mark the colon in each of the aforementioned chosen window settings using a dedicated, commercially available surface calculating software (Bluebeam, Pasadena, CA, USA). The slides were reviewed by all urologists under the same controlled viewing conditions. Consistent area measurement (the same reader marked the same area on both lung and abdominal window settings) was defined as an area calculation difference smaller than 0.1 cm².

OPEN IN VIEWER

FIG. 1.

Flowchart of study protocol.

Results

Ten urologists, four seniors and six residents, participated in interpretation of the 11 paired images as per study protocol, creating a study group of 110 comparisons (Fig. 1). Our patient cohort comprised eight men and three women. Six patients underwent left PCNL, and the other five patients underwent right PCNL. The mean delineated colonic area was significantly higher in lung window compared with the abdominal window with average of 8.82 cm² (±2.6) and 7.4 cm² (±2.2), respectively (p < 0.001). Switching the window setting from abdomen to a lung window resulted in detection of a larger colonic area in 55 cases (50%), no difference in 37 cases (34%), and smaller area 18 cases (16%). Figure 2 depicts how delineation was performed and a larger area in the lung window.

OPEN IN VIEWER

FIG. 2.

Differences in colonic area delineation between abdomen and lung window settings for a PCNL case. PCNL = percutaneous nephrolithotomy.

Intraclass correlation among all readers between delineated colonic area in each image was 0.87 (95% confidence interval [CI] 0.58–0.96, p < 0.001), whereas among seniors it was 0.92 (95% CI: 0.74–0.98, p < 0.001). Consistent area measurements between the window settings among seniors were noticed in 26/44 of the comparisons (59%), significantly higher compared with 17/66 (26%) of the residents (p = 0.008). There was no statistically significant difference in the colonic area identified using different window settings between patients scheduled for right or left PCNL.

Discussion

PCNL is a safe and effective technique to treat large renal stones. Despite the invasiveness of PCNL, its complication rates are low (3%–4%), with colonic injury being a rare complication. ^1,6 A meticulous examination of the patient’s preoperative imaging is important in avoiding this dreadful complication. Routine use of preoperative CT for planning proper percutaneous access has been suggested, ⁷ and NCCT with its cross-sectional imaging abilities has become the modality of choice for PCNL planning, and the American Urological Association and the European Urological Association support this practice. More attention should be paid to patients with previous operations, abnormal collecting systems, and mobile kidneys, which make access more difficult.

Different CT window settings have been investigated for the estimation of urinary stone size and density. ^8,9 A study by Danilovic et al. showed that kidney stone measurements vary according to NCCT window settings and that stone diameter and density differed among lung, bone, and soft-tissue windows. ⁸ Umbach et al. performed an evaluation of the accuracy of stone size measuring through NCCT in an in vitro study and found that measurements in all windows correlated significantly with the real stone size, but the bone window was more accurate. ⁹ In addition, several studies suggested that using a “lung-like” window during CT colonoscopy has a substantial influence on the visualization and correct size estimations of colonic polyps. The polyp size seems to be largest and most accurate (as compared with colonoscopy and pathology) when measured in the wider, lung window setting and less accurate when measured in soft-tissue windows. ^10,11 Consistently, in our study, larger areas of the colon were identified in the NCCT scans using the lung window, and this effect was more pronounced among residents.

To the best of our knowledge, our study is the first to evaluate the differences in colonic area measurements using different window settings before renal puncturing. We demonstrated that the lung window setting is superior in identifying the colon, and we believe that the lung window may be used by young urologists as an additional preoperative precaution for obtaining safe access to the kidney.

The main limitation of our study is the lack of the true anatomical colonic area to serve as the “gold standard” for comparison. Another limitation is the fact that we lacked information regarding body mass index of the patients. The impact of obesity on PCNL has been studied in several studies showing that normal-weight, obese, and super-obese individuals had similar outcomes. ¹² However, when assessing CT images, the amount of peritoneal and retroperitoneal fat may affect the reliability of our measurement.

Conclusion

Identification of the colon in multiple window settings, particularly in the lung window setting, provides the most accurate colonic demarcation for preoperative planning of PCNL access. This approach is more pronounced among young urologists and may reduce the risk of colonic injury.