Abstract
OBJECTIVE To investigate the optimized time period for detection and characterization of renal cell carcinomas (RCC) when the specific CT features appear during spiral dynamic CT scanning, and to optimize an effective scanning protocol of spiral CT for evaluating RCC.
METHODS Twenty-four patients with RCC verified by pathology had undergone a dynamic CT (D-CT) scan. A plain scan was employed to select the target slice. Single-level dynamic scanning started at 14-17 s after the intravenous contrast media had been administered, with a scan interval of 4.9 s acquiring a total number of 17~24 frames. A regular CT scan of the whole kidney followed by a delayed single slice acquisition through the target slice in the excretory phase was performed. Images were assessed in two ways: (1) A group of experienced radiologists reviewed the CT images to find when the specific signs appeared and when the CT features of RCC were optimally displayed; (2) Data measurement of the time-density curves (T-DC) of RCC. The exact time was obtained when the densities of the tumor, renal parenchyma, medulla and aorta reached their peak enhancement, thus also the time when the density difference between tumor and parenchyma was at maximum (Max T-M). Based on the slope of the contrast media uptake curve, T-DC types were ranked from the smallest to the biggest of slope as type A, B and C.
RESULTS 1. The review of the CT images by the radiologists showed that the CT features of RCC were optimally demonstrated at 70.2 s. The earliest time at which RCC CT features were examined was at 23.9 s. 2. Image data analysis: the time that the density (or CT value) of the tumor mass reached peak enhancement was at 54 s and peak value was at 80.4 Hu for RCC. The time of the maximal difference of densities between tumor and renal parenchyma was at 102 s.
CONCLUSION The following proposal is the scanning protocol for detecting RCC recommended by our research: After a plain scan to determine the target level, a single-level dynamic CT scan for renal tumors is to be programmed at scan delay time of 24, 38, 70,100 s, followed by a complete renal S-CT scan.
keywords
Bluem and Fishman[1] believe that the same strategy can be adopted for scanning both kidneys and liver. However, Yuh and Cohan [2] hold an opposing view based on the fact that the majority of abdominal spiral-CT scans are initiated at 40 ~ 70 s after the agent has been injected, the time when kidneys started to enter the advanced corticle phase. Although this delay provides the perfect time for examining other organs, such as liver, pancreas, spleen, etc., it is not suitable for exclusive renal check-ups. Up to now, little effort has been put into the systematic research of the time window for renal scans. Therefore our research has focused on the optimization of the time window for dynamic spiral-CT scans conducted within 2 min of a contrast media injection in renal cell carcinomas (RCC). Topics discussed include: (1) the earliest time when RCC signs appear and the time period when these signs are optimally demonstrated; (2) the strategy for obtaining the largest amount of RCC CT diagnostic information with a single injection of the contrast media and a relatively simple scan design. The study was designed in a prospective manner
Materials And Methods
Patients
Twenty-four consecutive adult patients with renal cell carcinomas (RCC) verified by histopathology, underwent a dynamic spiral-CT scan, and had their tumors surgically removed. Tumor histological specimens were produced from each case, and their one-to-one corresponding relationship with the CT information was analyzed for further study.
There were 16 men and eight women ranging in age from 13~79 years (mean, 50.5 years) and in weight from 46~65 kg (mean, 56.1kg). Clinical symptoms included gross hematuria with no pain, microscopic hematuria, severe or blunt waist pain or intermittent superior abdominal pain. In addition, three cases of renal masses were identified by coincidence through body examination.
None of the patients had a dysfunctional history of the heart, liver or kidneys.
Equipment and contrast media
A spiral-CT scanning instrument and contrast media, of 65% Angiografin or Ultravist (300 gl/ml, Shering CO. Germany), administered at a dose of 1.5 ml/kg at an injection rate of 2.0 ml/sec, were employed for all patients. The contrast media was administrated in a bolus via the cubital vein with a MEDRAD-OP 100 power injector. The onset of the delay time was specified immediately after the contrast media was injected.
Procedure design and scan technologies
Scans were performed following the steps listed below: (1) Patients were trained to breathe properly, with an abdominal belt to reduce respiratory artifacts. (2) Plain scan: the slice showing the largest area in a lesion was chosen. (3) Renal single-level dynamic spiral-CT scan: the single dynamic procedure was adopted. The scan was initiated 14~17 s after the contrast media had been injected, taking a total of 17 ~ 24 slices. The cycle time was 4.9 s for scanning frames with a scan slice thickness and slice gap of 8 mm, as well as a reconstructive slice thickness of 7.5 mm. (4) Complete renal spiral scan: After running the dynamic program, the Spiral program was immediately executed to initiate a complete renal spiral scan, with a scan bed speed of 7.5 mm/s, and a collimation of 7.5 mm, Pitch 1. (5) Excretory phase scan: Only the target slice was scanned, where the delay time varied in the range of 180~240 s.
Image post-processing
After scanning, the series of obtained images were further processed. First, a region of interest (ROI) in uniform size was selected on the tumor, renal cortex, renal medulla, renal pelvis, and abdominal aorta. Afterwards, the computer dynamic software was executed to process the image information acquired through dynamic CT (DCT). The time-density curves (T-DC) corresponding to the tumor, as well as the normal renal cortex, renal medulla, renal pelvis and abdominal aorta were automatically produced with time as the X-axis, and CT value as the Y-axis.
Principles for measuring lesion data: (1) Slices from which tumors, renal cortices, and renal medullas can be easily distinguished should be chosen, and a ROI is to be selected on the parenchyma of the tumor, the normal renal cortex, renal medulla, as well as the abdominal aorta. (2) Only one ROI needs to be chosen for homogeneously enhanced lesions. While for a non-homogeneously enhanced lesion, two ROIs should be selected on regions with a fairly different density. (3) For tumors with mixed densities, a parenchymal part that reflects the characteristics of the tumor should be chosen, while areas of cystoid change and necrosis, as well as the great vessels around or inside the tumor should be avoided to reduce partial volume effect. (4) The area of an ROI should be no larger than the region of a small tumor or the renal cortex, nor should it be too small (less than 3 mm2), such that CT values cannot be measured accurately. The area taken of ROI is usually 4 mm2.
Analysis of image data
Research was conducted based on visual review and data management.
Visual review
CT images were prospectively analyzed and reviewed using the visual impressions by two senior radiologists experienced in abdominal imaging. The features of single-level dynamic CT images were comparatively analyzed and examined until a unanimous conclusion was reached. The detailed procedure and research contents are as follows.
Normal renal enhancement phases
There were three phases of renal enhancement including a corticomedullary phase (CP), nephrographic phase (NP) and excretory phase (EP). Their CT criteria are outlined.[3]
Observation of abnormal CT features
These were identified in the first slice that possesses a density higher than that of the plain scan and displays abnormal CT features, as well as its delay time, which is considered as the time when symptoms start to appear.
The method and procedure to determine the “best” time for observing abnormal CT features are as follows: first of all, from the consecutive single-level CT images, the number of the types of abnormal CT features and the number of the slices that simultaneously display all types of CT features were verified; afterwards, the "best slice", that is the slice which presents the clearest image of all abnormal features (the middle slice was chosen in case of multiple best slices) was selected. In addition, the delay time of the “best slice” was also recorded.
Data management
A fixed ROI CT value (Hu) was applied to different regions for data management. The objective is to study how the CT values of the tumor, renal cortex, renal medullary, renal pelvis and abdominal aorta would change with time based on the time-density curves (T-DC).
Data management for tumors
The peak value and peak time of the tumor, renal cortex, renal medulla, renal pelvis, and abdominal aorta were recorded. Furthermore, the maximum difference between the CT values of each tumor case and that of the renal cortex plus renal medullary (maximum tumor-cortex, max T-C; maximum tumor-medulla, max T-M) was calculated. The detailed method is as follows: The Max T-C and Max T-M were obtained by first identifying the T-DC of the tumor, renal cortex and renal medulla, then computing the difference between the CT value of the lesion and that of the renal cortex as well as the renal medulla, and finally taking the absolute value of these differences.
The tumor T-DCs were categorized into three types based on their different slopes, namely type A, B, and C. The detailed method is as follows: first of all, the slope (S) of the rising sector of each curve, that is the maximum change in the curve’s CT value per unit time, was calculated. Afterwards, the curves were classified according to their S value into the above three types, specifically, Type A: 0<S<100 (Hu/sec); Type B: 100 ≤ S<200 (Hu/sec); and Type C: 200 (HU/sec) ≤ S. In addition, the peak value (PV) and peak time (PT) of each type of curves, along their average values were computed, wherein the PV and PT corresponding to different types of curves were further compared with variance analysis.
Results
Peer-review results
In the 17 ~ 24 consecutive single slices, slices revealing the earliest symptoms appeared within 18.7~38.5 s, with an average of 23.9 s, while signs were most optimally illustrated by slices produced in the time period of 35~131 sec, with an average of 70.2 sec.
Dynamic changes in ROI CT values
Fig. 1 demonstrates the T-DC for the abdominal aorta, renal cortex as well as renal medulla, and Table 1. shows both their PV and PT.
T-DC ofthe abdominal aorta,renal cortex,and renal medulla.
PV and PT of the abdominal aorta,renal cortex,and renal medulla.
Table 2.shows the PV and PT for RCC.
PV and PT of RCC.
Fig.2. shows the three types of tumor curves, and Table 3 illustrates the PV and PC for different types of RCC curves. With the exception of the PT for Type B and Type C (P=0.74, P>0.05), a significant difference was observed between the average PT and PV of all other types, wherein this difference is extremely significant between that of Type A and Type C (PT: P = 0.009, P < 0.001; PV: P = 0.0009, P < 0.0001).
Three types of tumor curves.
PT and PV of different types of RCC curves.
Table 4. shows the distribution of max |T-CI| and max |T-M|.
Distribution of max IT-CI and max IT-MI.
Shows the slice from which the earliest RCC symptoms were revealed. In the right renal tumor, the tumor vessels appeared in the shape of short branches with a high density, at the same time when the aorta image emerged.
Demonstrates the same case as Fig. 3, showing the slice where the RCC symptoms were most optimally revealed. Enhanced parenchyma part of the tumor in the corticle phase, combined with regions of cystoid change at a low density, wherein pseudo-capsules can be seen on the edges of the tumor.
Shows that in the excretory phase, the renal pelvis was filled with contrast media, while tumor parenchyma appeared at a low density.
Illustrates that the left renal cortex has not begun to enhance when RCC was in the CP. A “Vessel Pool” image was seen at the bottom left part of the left kidney, appearing at a time and enhancement density similar to that of the aorta.
Demonstrates the same case as Fig. 6, wherein the left renal cortex emerged at a later time comparing to that of the opposite side. The "Vessel Pool" image on the edges of the tumor remained at a relatively high density, while the circular enhancement of the tumor parenchyma and cystoid change in the shape of flakes were also present.
Shows the decrease in the density of the "Vessel Pool" image in the lesion.
Discussion
Based on the fact that no consensus has been reached on the extension of the time of renal CT scans, a more economic, convenient, and effective scan strategy is desired to optimize the examination of renal cell carcinomas. Therefore, aiming at gaining a thorough understanding of how different RCC signs vary with time under the spiral-CT condition, we conducted a study of single-level dynamic spiral-CT.
Factors affecting the time window of renal scans
The time windows of renal scans are inseparably linked to and affected by a number of parameters, including biological indicators, such as weight, heart rate, blood pressure, circulation time, functionality of the heart and kidneys, etc., as well as the type, dose, density, injection method, and injection rate of the contrast media.[3-6]In order to minimize the effect of the above factors, especially those of abnormal renal function, weight and contrast media on scan time windows, the studies of all cases included in this research were conducted on patients with normal heart and renal function, using a contrast media of 1.5 ml/kg. Furthermore, considering the middle-age and old patients among the studied objects, an injection rate of 2 ml/s was employed to reduce the possible occurrence of discomfort and contrast media leakage. This particular injection rate is commonly studied and utilized by researchers.[4,7-9]
The definition of renal phases and optimized temporal window
Based on the peer-review method, we believe that the slice from which the earliest RCC signs can be identified appears in the time frame of 18.7~37.5 s, with an average of 24 s. This slice primarily displays arteries that deliver blood to the tumor and their branches. Despite the fact that the time frame is fairly wide, its average value can still reflect the characteristic of tumors in the CP, as the tumors, are oversupplied by arteries.
Slices where symptoms are most optimally shown were discovered at a time between the CP and NP, specifically in the range of 35~131 s, with an average of 70.2 s. Since the tumor peak values also occurred during this time interval, we believe this time window can representingly reflect the typical characteristics of the enhancement of the tumor parenchyma. However, due to the inconsistency between the circulation time of renal cortices and renal medullae, non-entirely enhanced renal cortices are easily confused with small-enhanced masses in this unstable period.[10-14] Therefore, it is reasonable to consider that in order to eliminate pseudo-positive and pseudo-negative results, this time window should not be used as the only analytical reference in tumor diagnoses. Among the tumor cases studied in this research, one (one case of cystoid renal cell carcinoma) demonstrated delayed enhancements, with peak time at approximately 100 sec. Based on this finding, the significance of scans with a 100 s delay time in the diagnoses of delayed enhanced tumors is thus especially important.
With reference to the results illustrated in this report, the average value of all tumor PTs, i.e. 54 s, along with the ones for RCC, were not considered as having statistical significance. In addition, for the purpose of obtaining a better assessment of the trend of tumor parenchyma enhancement curves during 14-60 s, tumors were classified into three types, namely Type A, B, and C, according to the slope of the rising sector of their T-DCs. In this regard, Type A represented tumors with the minimum density change per unit time, Type C symbolized those with the maximum change, while Type B falls between the other two types. As shown in the results, the average PT values for Type B and C were 53 s and 52 s respectively, leading to a difference that does not possess any statistical significance. Furthermore, 70% of all tumors demonstrated their maximum change in density during this period. This time period should therefore be included in the optimized scan time window for observing tumor enhancement, and, especially, the characterization of small masses. In addition, because changes in the tumor CT values gradually grew steady after the peak time, which almost matched with the occurrence of the slice presenting the most optimal image of the symptoms, identical tumor CT values were shown on two consecutive slices. Therefore, a single one can be chosen between the two consecutive slices.
When the renal cortex enhanced to a certain degree at the tumor peak value, the enhancement of the renal medulla was also initiated. This is why the ratio of tumor to renal cortex did not always remain at the maximum value. Furthermore, the max T-C had appeared in two different time slots, either in the early stage of the CP or in the NP, adding to the uncertainty in the scan time window for tumor examination. The max T-M was evenly distributed within 102 s towards the end of the curves, suggesting that the ratio of the tumor density to that of the renal cortex was at a maximum and that the tumor lesion was most optimally illustrated in this time window. Based on the above facts, this particular time period should be included in the scan time as well. Small-scale (diameter ≤ 1.5 cm) renal pelvis carcinomas were not covered in the scope of this research.
Based on the above tumor characterization and emergence indexes, the authors of this report believe that scans performed 14~120 s after the contrast media has been injected can best reflect the indicators for evaluating tumors, including the time for the slices of the earliest and the most optimally revealed symptoms, as well as the time of the largest difference in the tumor density and that of the renal medulla. The first two indications have possessed a greater value for tumor characterization, while the latter one contributes more significantly to tumor detection. In order to maximize spiral-CT function in tumor characterization and detection, it is suggested that single-level scans be conducted with a delay time of 24 s, 38 s, 70 s, and 100 s, followed by a complete renal scan (EP), as well as a target slice scan after waiting another 180 s. The first two single slices are intended to illustrate the early symptoms of the tumor; the third one is for presenting the most optimal image of the symptoms, while the fourth one displays the biggest contrast in the density of the tumor and its surrounding renal parenchyma. The delay time may vary with different examination objectives.
This research suggests a method where only a single contrast media injection is needed for characterizing and detecting tumors. Specifically, an ultrafast spiral-CT is first conducted to select a target slice via a plain scan, followed by single-level dynamic multi-phase enhancement CT scans with appropriate contrast media injection technology, and a complete renal scan. By minimizing the amount of the contrast media required as well as exposure, the best cost-effective ratio for tumor diagnoses can be achieved, thus realizing the practical values of the proposed scan strategy.
- Received January 30, 2005.
- Accepted March 22, 2005.
- Copyright © 2005 by Tianjin Medical University Cancer Institute & Hospital and Springer
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.