Efficacy of transrectal ultrasonography (TRUS) in preoperative staging of rectal cancer ultrasonografije (TRUS) u preoperativnoj proceni stadijuma rektalnog

Background/Aim. The outcome of rectal cancer is dependant on the stage of the tumour. There are several classification systems used to describe the extent of the disease. The aim of this study was to compare the efficacy of transrectal ultrasonography (TRUS) in preoperative local staging of rectal cancer using endosonographic probes with different views (180° vs 360°), as well as an influence of experience of an endoscopist on the TRUS performance. Methods. TRUS was performed in 127 patients with rectal carcinoma by two endo-scopists. Seventy-one patients were examined with a 180° endosonographic probe (group A) and 56 patients with a 360° rotating probe (group B). All findings were compared with a histopathology report. Results . TRUS had a diagnostic overall accuracy of 91.3% for the tumor (T) category (k = 0.866, SE (k) = 0.038, p < 0.0001) and 71.7% for the node (N) category ( κ = 0.374, SE (k) = 0.082, p < 0.0001). In the group A, TRUS had a diagnostic overall accuracy of 88.7% for the T category ( κ = 0.805, SE (k) = 0.063, p < 0.0001), and 70.4% for the N category ( κ = 0.376, SE (k) = 0.101, P < 0.0001). In the group B, TRUS had a diagnostic overall accuracy of 94.6% for the T category ( κ = 0.920, SE (k) = 0.044, p < 0.0001), and 73.2% for the N category ( κ = 0.379, SE (k) = 0.131, p = 0.004). Experience of the endoscopist had no significant influence on results of preoperative staging of rectal cancer by using TRUS. Conclusion . The accuracy rate of TRUS in the preoperative local staging of rectal cancer is high. Our results imply no significant difference in the overall accuracy rates when using endosonographic probes with different views (180° vs 360°). Also, there was no significant influence of endoscopist experience on results obtained. histopathologic using the different


Introduction
Colorectal cancer is the third most common cancer in Europe and the USA, and the third most common cause of cancer related deaths. Over 50% of patients have locally advanced disease that has spread to the lymph nodes and/or the liver at the time of diagnosis 1, 2 . The outcome of rectal cancer is dependant on the stage of the tumour. There are several classification systems used to describe the extent of disease. In this study, transrectal ultrasonography (TRUS) tumor stage was assessed by the Tumor-Node-Metastasis (TNM) classification as described by Hildebrandt and Pfeifel 3 .
The managment of rectal cancer has evolved to become multidisciplinary because it offers the best clinical outcome, although surgery remains the most important treatment 4, 5 . This greatly increased the importance of the accurate preoperative staging in providing information about tumor infiltration and lymph node metastasis in order to make the right decision regarding rectal cancer treatment.
TRUS introduced by Wild and Reid in 1956, is very accurate imaging modality for the assessment of tumour growth in the bowel wall with the reported overall accuracies for the T and N staging between 69%-97% and 58%-83%, respectively [6][7][8][9] . Moreover, TRUS is inexpensive and quick diagnostic procedure associated with minimal discomfort to the patient.
The TRUS probes exist as radial and curved linear array depending on the orientation of the ultrasound transducer. The radial probes produce a 360° picture in a plane vertical to the long axis of the endoscope insertion tube, while a linear array create sector-shaped images horizontal to the long axis of the insertion tube 10 . Assessment of the wall of rectum and nearby structures is best achieved with radial probes with a frequency ranging from 6-16 MHz. Within these probes, two crystals are attached back to back, and can rotate inside the transducer 11 .
The aim of the present study was to determine the accuracy of TRUS in rectal cancer staging compared with a histopathologic examination using the rotating endosonographic probes with different views (180° vs. 360°), and to evaluate the influence of experience of an endoscopist on the TRUS performance.

Methods
Preoperative TRUS was performed in all patients presented to the Clinic of Gastroenterology, Clinical Centre of Serbia, Belgrade with newly diagnosed rectal cancer who had no previous tumor staging evaluation. Patients with previously performed staging (MRI of the pelvis) were excluded. During 6-year period, 127 TRUS examinations were performed for the staging of rectal cancer by two endoscopists. Seventy-one TRUS examinations were conducted using a biplane endorectal probe with a field of view of 180° (Hitachi EUB 6500 U533), while 56 TRUS examinations were performed using the endorectal probe with a full 360° field of view (BK medical 1850). As the operator physically move the 1850 probe while the transducer moves along the entire length of the tumor and provides an image in the axial direction, the U533 biplane probe provides information both axial and sagittal.
The patient selection regarding the technique of TRUS was performed according to the department where they presented first. Informed consents were obtained from all of the patients prior to the examination. Before the probe was inserted into the rectum, a digital rectal examination was carried-out to identify the size, fixation, morphology and location of the tumor and to exclude clinically important stenosis to determine whether the anal canal and lower rectum are passable. All patients were evaluated to determine the diagnostic accuracy of depth of transmural tumor invasion and lymph node metastases. The TRUS results were correlated with the histopathological reports regarded as the gold standard in local staging of rectal carcinoma.
The TRUS T stage was assessed by visualising the depth of the tumour penetration through five defined layers of echogenicity in the rectal wall as described by Hildebrandt et al. 12 . All identified lymph nodes were measured and nodes greater than 5 mm in the maximum diameter were classified as positive (N+). The nodes smaller than this were assumed to be normal or inflammatory and were defined as N0. Comparison was made between the ultrasound staging and histopathologic findings after surgery.
Statistical analysis was performed using the Measure of Agreement-Kappa test for accuracy rates of T and N staging. Comparison of the accuracies within both the T and N staging results was made using the Fishers Exact Test or χ 2 -test, with a p value of < 0.05 considered to be significant.
In order to determine the influence of experience of the endoscopist on the TRUS performance, TRUS performed during this period was divided into two time periods. The first time period was taken as the first half of practice, and the second period was taken as the second half of practice. Accuracy of T-and N-staging was calculated and compared in each time period.

Results
The total of 127 patients were examined by TRUS (90 males and 37 females, median age 63 years, range 26-85 years), and all of them underwent surgery. After surgery, preoperative findings were compared with histopathology findings of the surgical specimen.

T -tumor; N -node.
For the purpose of our analysis, the patients were divided into two groups. First group was examined with a 180° rotating endosonographic probe (group A, 71 patients) and the second group was examined with a 360° rotating endosonographic probe (Group B, 56 patients).
A high accuracy rate was maintained throughout the study period for the T staging in both groups. There was a slight improvement in the accuracy rate of the T staging in the group A from 83.3% in the first half of practice to 97.1% in the second half of practice, although the difference was not statistically significant. In the group B, a high level of accuracy in the T staging were maintained throughout the study -from 92.9% in the first half of practice to 100% in the second half of practice. There was a decrease in the accuracy rate of the N staging in the group A from 81% in the first half of practice to 60% in the second half of practice, although the difference was not statistically significant. In the group B, a high level of accuracy in the N staging was maintained throughout the study -from 68% in the first half of practice to 79% in the second half of practice.

Discussion
At present, a combination of computed tomography (CT), magnetic resonance imaging (MRI) and TRUS, is used for the preoperative staging of rectal cancer. A choice of modality depends on local expertise and availability.
For assessing the depth of tumour growth in the bowel wall, TRUS is very accurate with reported overall accuracies for the T staging varying between 69% and 97% 9 . On the other hand, CT is the current standard for staging of distant metastasis and cannot be considered appropriate for the local tumor staging 13 . MRI seems to be superior for more locally advanced disease with reported sensitivity between 66% and 92% 14 . Two meta-analyses showed that sensitivity was affected by the T stage 15,16 . TRUS seems to be more accurate for staging of superficial rectal T1 and T2 tumours, with reported sensitivity of 94%. A report of a large endosonography study in 1,184 patients with rectal tumors confirmed these findings with the overall staging accuracy of 69% that is lower than previously reported because of less accurate assessment of the local tumor extent in advanced rectal cancer 17 . On the other hand, study conducted in Israel reported that the accuracy of TRUS in the local tumor staging was more accurate for T1 (81.2%) and T3 (94.1%) in comparison with T2 (63.6%) 18 .
In our study, the overall accuracy rate in determining the depth of tumor invasion was 91.3%. Accuracy rates for T1 and T2 tumours were 88.9% and 91.9%, respectively. The highest accuracy rate was for T3 (93.3%). Overstaging was found in 4.7%, and understaging in 3.9% of 127 patients. Thus, our results are comparable to those reported in relevant literature 19 . A reason for good results of this study is the level of experience of the endoscopists. Both operators were highly experienced endosonographers that demonstrated superior performance, underscoring the existing learning curve for mastering endoscopic ultrasonography. The improvement with experience was shown by Orrom et al. 20 , who found that the staging accuracy of rectal cancer increased from 58% in the initial 12 examinations to 88% for the subsequent 24 procedures. In our study, a high levels of accuracy in the T staging were maintained throughout the study in both groups -from 83.3% in the first half of practice to 97.1% in the second half of practice, and from 92.9% to 100%, respectively.
Assessment of pararectal lymph node involvement is essential for a selection of a high risk patients which are candidates for preoperative chemoradiotherapy, and still represents a diagnostic problem. Meta-analysis of 6 included studies showed that TRUS was only slightly superior to noncontrast enhanced MRI and CT in identifying lymph node metastasis with reported accuracy rate from 58%-83% 8,21 . CT cannot accurately distinguish between malignant and benign lymph nodes with nodal staging accuracy between 54% and 70%. The MRI accuracy was found to range from 60% to 90% for lymph node metastases 14,[22][23][24][25] . In our study, the overall accuracy rate of assessing a lymph node status was 71.7% which was similar to the previously reported results. There was no significant difference between the groups in the overall accuracy rates of assessing the lymph node status. It seems that 360° view is not superior to 180° view in better visualization of perirectal lymph nodes.
A high accuracy rate for the N staging in this study (with a cut-off of 5 mm for positive nodes) was somewhat surprising as we were aware that almost 30-40% of the involved nodes were of 4 mm diameter or less. However, this should be viewed through the prism of a high level of false negative and false positive rates reported in the study. There was a tendency for overstaging nodes in both groups.
We are aware that this study has potential drawbacks. Only the patients without previous staging were included in the study, so this could be a source of selection bias. A lack of randomization is the most important drawback, since patients were not randomized for the technique of TRUS. Although it may be a potential source of error, we believe that this issue could not significantly influence results since the patients were not deliberately selected, as the type of TRUS was determined according to the unit where a patient was first presented.

Conclusion
In conclusion, the accuracy rate of TRUS in the preoperative local staging of the rectal carcinoma and regional lymph node involvement is high. Our results imply no significant difference in the overall accuracy rates of assessing local and lymph node status when using the endosonographic probes with different views (180° vs 360°) with an exception of accuracy in the T1 staging where 360° was superior to 180°.