Abstract
OBJECTIVE To prospectively investigate the correlation between the enhancement parameters of a dynamic-CT (D-CT) scan for renal cell carcinomas (RCC) and the carcinoma tissue microvessel density (MVD) in renal cell carcinomas (RCC).
METHODS Twenty-four cases of renal cell carcinoma verifyied by histopathology were scanned via dynamic-CT, followed by a whole kidney scan. Enhancement parameters were derived as follows. The slope of the contrast media uptake curve (S), area under the curve(AR), the density difference before and after tissue enhancement (Δ HU) and tissue blood ratio (TBR) were calculated for all lesions. Time-density curve types were ranked from the lowest to the highest of the slope of the contrast media uptake curve (S) as type A, B and C. Pathologic slides corresponding to the CT imagings were subjected to CD34 monoclonal antibodies, then were evaluated with an image analyzer to count hot spots of MVD. By using the Spearman rank correlation tests, statistical analysis was performed to determine the strength of the relationship between enhancement parameters and MVD determinations.
RESULTS The carcinoma tissue MVD showed a direct correlation with the enhancement parameters of D-CT (r=0.54, r=0.62, r=0.55, r=0.64, r=0.44, P< 0.05). Moreover the S, Δ HU, TBR and type curves all demonstrated a strong correlation with the MVD. By analyzing the various enhancement parameters of the time-density curves, the relationship between the enhancement CT parameters corresponding to the tumor's MVD was identified.
CONCLUSION A dynamic spiral-CT scan may be a helpful method as a measurement of tumor angiogenesis in vivo in RCC.
keywords
Tumor angiogenesis is a process in which blood capillaries emerge from tumors and/or surrounding tissues, and vessel growth is commonly measured based on tumor microvessel density (MVD). In various multiple clinical series and for multiple tumor types, including breast, lung, prostate, and head and neck carcinomas, MVD has been shown to be a useful and independent indicator of prognosis. Anti-angiogenesis has already become a new strategy for treating tumors, due to its evident effectiveness in constraining tumor growth. [1-3] However, our understanding of the processes of tumor angiogenesis is far from complete. The histological capillary density technique to characterize tumor angiogenesis, although currently a favored method, may not be ideal because it is invasive and subject to variability in interpretation and random tumor sampling. With major advances in medical imaging, in particular for analysis of renal cell carcinoma, we may readily obtain high-resolution angiogenic data in addition to volumetric quantitative information on tumors. Growth and the temporal course of a tumor can be followed more closely coupled to the angiogenic and vascular progression of the tumor. Angiogenic data will provide dynamic validity to current methods of analyzing tumor vascularity, such as immunohistochemical analysis. CT guided evaluation for the tumor vasculature can help decipher the role of current, and to be developed, strategies targeting aberrant angiogenesis in renal cell carcinoma. The purpose of this study was to analyze the correlation between dynamic spiral-CT enhancement parameters and tumor angiogenesis in renal cell carcinoma in search of a noninvasive method for assessing tumor angiogenesis in vivo.
Materials and Methods
Patients
During the period from October 2000 to March 2004, renal cell carcinomas (RCC) verifyied by histopathology had undergone a dynamic-CT (DCT) scan in 24 consecutive patients (16 males and eight females; mean age, 50.5 ± 16.7 years; age range, 13~79 years). Tumor histological specimens were produced from each case, and their one-to-one corresponding relationship with the CT information was analyzed for possible correlateions. All patients described above had not been seriously affected by malfunctioning of the heart, liver, or kidneys. The study was designed in a prospective manner and was approved by the Institutional Review Board. Informed consent was obtained from all patients.
Equipment and contrast agents
DCT was performed (Samaton plus 4; Siemens, Erlangen, Germany). All patients received one injection of Ultravist 300 (Schering, Berlin, Germany) (300 mg of iodine per milliliter given per 1.5 ml of body weight). In all patients, the contrast medium was injected with a rate of 3 ml/sec through a cannula that was placed in the forearm. A CT power injector (Envision; Medrad, Indianola, PA) was used for the injections.
Procedure design and scan technologies
Scans were performed through the following steps: I) Patients were trained for proper breathing, and fastened with an abdominal belt to reduce respiratory artifacts. 2) Plain scans: identified the largest renal lesion slice. 3) Renal single dynamic CT scan: the single dynamic procedure was adopted. The delay time was 14 ~ 17 s after the contrast agent had been injected and 17-24 slices were scanned. The cycle time was 4.9 s. 4) Whole kidney scans: After running the sequence procedure, the spiral program was immediately executed to initiate a whole kidney scan. 5) Last slice scan: Only the target slice was scanned at last.
Image post-processing
First, a region of interest (ROI) of the same size was selected on the tumor and abdominal aorta for measuring their CT values. Afterwards, computer software was executed to process the image information acquired through DCT. A corresponding time-density curve (T-DC) was automatically produced.
Measurement principles of lesions: 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, and the abdominal aorta. 2) Only one ROI needs to be assigned 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 varying densities, a parenchyma portion that reflects the characteristics of the tumor should be chosen, while areas of cystoid change and necrosis, as well as the great vessels around and inside the tumor should be avoided. 4) The area of an ROI should not be so large as to exceed that of a small tumor or the renal cortex region, nor should it be too small (less than 3 mm2), such that CT values cannot be measured accurately. The area taken is usually 4 mm2.
Analysis of imaging data
The density difference before and after tissue enhancement (ΔHU) = maximum CT value (peak value) after tissue enhancement — CT value acquired via tissue plain scan.
The slope (S) of the time-density curve = (peak value after tissue enhancement —CT value before tissue enhancement) / Time period of the tissue peak value.
Tumor T-DC curves were divided into three types based on their different slopes, namely type A, B, and C. Type A: 0<S<100 (HU/sec); Type B: 100≤ S<200 (HU/sec); and Type C: S ≥ 200 (HU/sec).
Tissue blood ratio (TBR) = (peak value after tumor tissue enhancement —CT value before tumor tissue enhancement) × 100% / (peak value after aorta enhancement —CT value before aorta enhancement).
Areas (AR) under the T-DC that correspond to different tissues were automatically calculated with the D-CT program.
Immunohistochemical staining of tissues
The analysis and observation of pathologic specimens: tumors were excised. Transverse cryosections (4m) were obtained through the middle part of the tumors, corresponding to the transverse orientation of the selected CT imaging sections. The cryosections were immunohistochemically stained by using monoclonal antibodies (CD34) (Zhongshan Biological Preparation Co., Ltd, Beijing, China). In addition, five samples of non-tumor kidney tissues were extracted from surgically removed adult kidney specimens to serve as comparative reference.
Analysis of the immunohistochemical results
MVD values were calculated from the mean number of vessels of the tumor rim, the tumor core and microvascular hot spots[4] (i.e., the five highest MVD values per tumor cryosection) in a microscopic field of 0.2 mm2(magnification × 200).
Statistical processing
Quantitative data were expressed with mean ± SD, and analyzed based on variance analyze; In addition, Spearman rank correlation coefficients were used to assess the correlation between each of CT imaging and tumor angiogenesis. The Spearman rank correlation allows statistical inference from a non-normal distribution of variables. Statistical analyses were performed by using commercial software (SPSS 10.0.5; SPSS, Chicago, III), and P<0.05 was considered to indicate a statistically significant difference.
Results
Immunohistochemical results
CD34 staining
Microvessel distribution and shape: The shape, distribution, and quantity of capillaries: Capillaries vary greatly in shape, where they often scatter or diffuse in the shape of a rift, a net, and a tube. Capillaries are unevenly distributed in carcinoma tissues, specifically dense on the edges of the tumor, the region where carcinoma cells reproduce most actively, and are scarce or even nonexistent in the central part or close to the necrosis areas of the tumor. The average MVD for RCC was (148 ± 27)/0.2 mm2.
Normal renal tissues
Capillaries, with an average MVD of (98.5± 20.5)/0.2 mm2, appear and are distributed in a relatively more homogeneous and orderly array.
After analyzing and comparing the two by variance, a major difference was observed between the average MVD of renal cell carcinomas (RCC) and that of normal renal tissues (P=0.0002), suggesting the MVD of RCC is significantly higher than that of normal renal tissues.
Dynamic CT enhancement parameters and their relationship with MVD
Classification of the time-density curves (T-DC) for tumors (Fig. 1).
Graph shows three time-density curves of renal cell carcinoma.
Dynamic CT enhancement parameters, MVD values (Table 1, Figs. 2,3).
Renal cell carcinoma in a 53-year old man. A: Demonstrates the circular enhancement and the “blood pool” image of the tumor parenchyma in the left side mixed RCC renal cortex phase. B: CD34 staining (enlarged 200 times), densely distributed tumor capillaries.
Renal cell carcinoma in a 48-year old man. A: Demonstrates a small renal carcinoma. Tumors appear with a relatively low density in the renal parenchyma phase. B: Demonstrates the same case. CD34 staining (magnification x 200), tumor capillaries appear in pale-brown and distribute in the shape of sprouts.
DISCUSSION
Angiogenesis occurs physiologically during embryonic development, in the female reproductive cycle, wound healing, and hair growth. The primitive vascular network is modified via angiogenesis, which leads to maturation, branching and formation of a complex vasculature that is seen in adult life.
The analysis of dynamic CT imaging: The slope (S) of the contrast media uptake curve of the tumor T-DC reflects the value that enhances the rapid density change in early-stage tumors. The current study combining imaging with histopathology indicated that the tumor early-stage enhancement rate correlates with the quantity of vessels, and tumor vascularity can partially explain the swift change in the distribution of the contrast agents in malignant tumors. [5-7] In addition, the result in our study has also revealed a highly significant correlation between S and MVD. It is consistent with the conclusion drawn from the previously mentioned study, suggesting that the slope of the contrast media uptake curve is in fact an excellent indicator for tumor angiogenesis. Furthermore, tumor enhancement values (ΔHU) and tissue blood rates (TBR) are generally employed to reflect tumor enhancement by their density change after a contrast media has been injected. Researches have shown that the injected contrast media tends to flow in tissues with densely populated vessels, and extracellular space, thereby demonstrating that angiogenesis is the foundation for tumor enhancement.[8,9] This study has further pointed out the significant and direct correlation between MVD and ΔHU, TBR, suggesting that lesion enhancement will become more intense as the number of tumor vessels increases. However, scholars have proposed besides tumor angiogenesis, lesion enhancement is also affected by other factors, such as the perfusion rate within tumor vessels and the size of the extracellular space. [10] Although the ΔHU value of three types of RCC does vary significantly, one should consider the weight and body fluid of each individual patient. Therefore the use of ΔHU as an indicator of tissue enhancement is rather approximate.[11]
Aiming at eliminating the above factors, TBR values were adopted by this study to indicate the maximum enhancement of the tumor tissues objectively and precisely. When comparing the TBR values of three types of tumor curves, the correlation test results are clearly different from that of ΔHU, although TBR isstrongly related to MVD as well. The reason for this might be the huge difference between tumor and abdominal aorta enhancement, leading to an enhancement ratio that is small enough to disguise the difference between the TBR values for various curve types. In addition, the area under the time-density curve is the summation of the area under the rising sector, that is the rapid enhancement phase, and the area under the sector after the peak value, also known as the delayed enhancement phase. The parameters (Δ HU, S, TBR) of the rising sector were found to relate to tumor angiogenesis, thus, theoretically, such correlation should also be present in the area under the curve.
The data included in this research have illustrated that AR is directly proportional to MVD; specifically MVD would increase if AR increased. It is also possible that AR is an overall reflection of a number of factors, including the perfusion rate of iodine in tissues, how long it is detained in tissues, etc. The biological meaning of AR needs to be examined and researched further.
The RCC carcinoma tissue MVD corresponding to three curve types: As revealed by the data included in this research, the RCC carcinoma tissue MVD corresponding to three curve types differ from one another in the mean value of the tumor MVD, wherein this difference tended to be distinct between that of type A and type C, though there was no statistical significance. It may be apparent that there are differences between the angiogenesis of RCC carcinoma tissues of the same cytological type. For the current case, type A may tend to symbolize tumors of light angiogenesis, while tumors of intensive angiogenesis may be represented by type C. These differences may be a result of the variation in the effectiveness of angiogenic control and interventional therapy in treating tumors of the same cytological type. More cases should be obtained to explore this result.
By analyzing the data included in this research, it is evident that dynamic CT parameters, including S, Δ H U, AR, TYPE and TBR, are directly proportional to MVD. This is to say that the quantity of tumor tissue capillaries can be preliminarily assessed in vivo via the analysis of data from the tumor CT, the observation of the dynamic enhancement process, and the analysis of related parameters as well as indexes. However, in the future, CT guided evaluation of angiogenesis should be conducted comparing benign with malignant masses.
In summary, the in vivo assessment of tumor angiogenesis in renal cell carcinoma tissues via spiral-CT provides a crucial base for choosing tumor treatment schemes, such as gene therapy for tumor angiogenesis control, interventional therapy for tumor vessel embolism, and surgical removal, as well as predicting the prognoses of tumors.[12-14]
ACKNOWLEDGEMENTS
We thank Mr. Li Ning for technical assistance, photography and typing the manuscript.
- Received January 22, 2005.
- Accepted March 10, 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.