| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Genitourinary Imaging |
1 From the Departments of Radiology (Y.Y., T.B., Y.B., R.N., S.I., M.T.) and Obstetrics and Gynecology (H. Ohtake, H. Okamura), Kumamoto University School of Medicine, 1-1-1 Honjo, Kumamoto 860-0811, Japan. Received July 8, 1999; revision requested August 25; revision received January 14, 2000; accepted January 27. Address correspondence to Y.Y. (e-mail: yama@kaiju.medic.kumamoto-u.ac.jp).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONAPPENDIXREFERENCES |
|---|
MATERIALS AND METHODS: Dynamic enhanced MR imaging and pharmacokinetic analyses were performed in 26 patients with cervical cancer who subsequently underwent hysterectomy and in 36 patients with cervical cancer who received radiation therapy. Histopathologic findings and clinical outcomes were correlated with results of dynamic MR imaging and pharmacokinetic analysis.
RESULTS: On dynamic MR images of the surgical patients, areas with intense homogenous enhancement showed increased permeability (k = 27.4 x 10-3) compared with areas with poor enhancement (k = 19.0 x 10-3). Well-enhanced areas were predominantly composed of cancer cell fascicles, whereas poorly enhanced areas were composed of fibrous tissue with scattered cancer cells. Radiation therapy was more effective in tumors with higher tissue permeability (k = 31.3 x 10-3) on dynamic MR images than in those with lower tissue permeability (k = 18.3 x 10-3).
CONCLUSION: Areas of increased contrast enhancement are mainly composed of abundant cancer cell fascicles, whereas poorly perfused areas are composed of fibrous tissue with scattered cancer cells. Radiation therapy is more effective in well-enhanced tumors, resulting in improved local control.
Index terms: Magnetic resonance (MR), contrast enhancement, 854.12143 • Magnetic resonance (MR), treatment planning, 854.32, 854.1299 • Uterine neoplasms, MR, 854.121411, 854.12143 • Uterine neoplasms, therapeutic radiology, 854.32, 854.1299 • Uterus, surgery, 854.32
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONAPPENDIXREFERENCES |
|---|
Radiation therapy remains the treatment of choice for advanced cancer of the cervix. However, some patients with this cancer cannot be cured with conventional radiation therapy alone (7,8). The addition of chemotherapy and/or radical surgery may further improve patient outcome, but they have not been routinely used because of their increased risk of complications (9–11). Traditional prognostic factors, including clinical staging, tumor size, and tumor extent, may not always allow effective prediction of the outcome. To our knowledge, there are no well-established means by which the subset of patients who will be resistant to radiation therapy can be identified before therapy.
Dynamic enhanced study enables us to evaluate the microcirculation of tumors (4,9,12). The finding of hypovascularity may suggest poor-oxygenation status, which has been considered to be one of the multifactorial causes of resistance to tumor treatment. Techniques that can be used to accurately, rapidly, and noninvasively assess tumor vascularity and oxygenation may, therefore, have the potential to allow more effective prediction of the response to radiation therapy and to allow early identification of those patients who may benefit from the addition of novel and/or more aggressive therapies. Although there are ample reports on dynamic enhanced study of cervical cancers, the histopathologic basis of the enhancement and its importance in clinical outcome have not been fully understood.
The purpose of this study was to investigate the histopathologic bases of different enhancement patterns on dynamic MR images by correlating imaging findings with histopathologic and pharmacokinetic results and to assess their importance in the prediction of outcomes in patients with advanced cervical cancer.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONAPPENDIXREFERENCES |
|---|
Twenty-six women (age range, 27–75 years; mean age, 52.7 years ± 5 [SD]) with cervical cancer underwent both dynamic MR imaging and surgery. To qualify for this study, patients had to have a histopathologic diagnosis of cervical carcinoma classified with clinical criteria as stage IB or higher (IB, n = 10; IIA, n = 4; IIB, n = 12), and whole pathologic specimens had to be available. In all patients, MR imaging was performed within 2 weeks of surgery. Among the 26 patients, 20 had squamous cell carcinoma, and six had adenocarcinoma of the cervix. Six tumors were well differentiated, 10 were moderately differentiated, and 10 were poorly differentiated. The volume of tumor measured on MR images ranged from 16 to 212 cm3. Radical hysterectomy with pelvic lymph node dissection was performed in 10 patients, and pelvic exenteration was performed in 16 patients.
All MR findings were correlated with the histopathologic findings because every cervical carcinoma, irrespective of stage, was surgically removed. Histopathologic examination was performed in all patients by using whole specimens in planes that resembled those of the MR imaging sections and by using standard histomorphometric techniques, including hematoxylin-eosin, Azan, and factor VIII staining. The enhancement pattern on dynamic MR images was correlated with the amount of tumor cells, interstitial fibrous tissue, and vessels.
Correlation of Dynamic MR Imaging Findings with Results of Radiation Therapy
Dynamic MR imaging and pharmacokinetic analyses were performed before primary radiation therapy in 36 patients with cervical cancers, which included those of stages bulky IB (n = 2), IIA (n = 1), IIB (n = 8), IIIA (n = 3), IIIB (n = 14), and IVA (n = 8). Initial clinical assessment of tumor stage was performed by a gynecologic oncologist (H. Okamura). Pretreatment clinical evaluations in all patients included examination at chest radiography and chest, abdominal, and pelvic computed tomography (CT). Clinical evaluations performed during radiation therapy included joint examinations by the gynecologic oncologist and a radiation oncologist (T.B., Y.B., R.N.) to assess clinical tumor response.
Among the 36 patients (age range, 30–83 years; mean age 56.2 years ± 4.7), 33 had squamous cell carcinoma, and three had adenocarcinoma of the cervix. Ten tumors were well-differentiated, 11 were moderately differentiated, and 15 were poorly differentiated. The volume of tumor ranged from 5 to 236 cm3 on MR images. There were no statistically significance differences between the surgical and radiation groups in patient’s age, tumor volume, histologic subtype, and histologic grade, although the radiation group included patients with cancers of more advanced stages.
Median follow-up was 30 months (range, 6–87 months). The doses of external beam radiation ranged from 45.0 to 61.2 Gy. Intracavitary therapy was added to the external beam therapy in 29 patients. The MR imaging protocol and method of pharmacokinetic analysis in the radiation group were similar to those of the surgical group. Follow-up clinical evaluation included a pelvic examination performed by the gynecologic oncologist. Pap smears were performed at 4–6-week intervals for the first 3 months after completion of treatment and at 2–3-month intervals thereafter.
MR Imaging Protocol
MR imaging was performed by using 1.5-T superconducting units (Magnetom Vision, Magnetom Symphony; Siemens, Erlangen, Germany) with a consistent imaging protocol. MR imaging was performed at least 3 weeks after biopsy and within 7 days before surgery. Transverse or sagittal T1-weighted spin-echo images were obtained before and after contrast enhancement with gadopentetate dimeglumine (Magnevist; Schering, Berlin, Germany) with a repetition time msec/echo time msec of 600/15, a 256 x 512 matrix, and two signals acquired. Transverse and sagittal nonenhanced T2-weighted images were obtained with a turbo spin-echo sequence with 4,500/120, an echo train length of 15, a 300 x 512 matrix, and two signals acquired. To evaluate the perfusion of the tumor, dynamic enhanced imaging was performed after bolus injection of gadopentetate dimeglumine by using a turbo spin-echo technique with 340/12, an echo train length of three, a 128 x 256 matrix, and one signal acquired. Gadopentetate dimeglumine was intravenously administered at a dose of 0.1 mmol per kilogram of body weight. The injection was administered at a flow rate of 1 mL/sec, and scanning was started at the end of the injection. Actual sampling time for a single dynamic acquisition was 30 seconds, and five acquisitions were obtained with both MR units to evaluate contrast enhancement. Images were obtained in the transverse or coronal plane with a section thickness of 5–7 mm and 1.4-mm intersection spacing.
Analysis of Dynamic Contrast-enhanced Images
At quantitative analysis of each MR imaging examination, the region of interest (ROI) was chosen on the basis of best tumor delineation on both nonenhanced and enhanced images. In each lesion, three 4–10-mm2 ROIs that were representative of the tumor were selected by one of the authors (T.B.), who did not know the histopathologic or clinical results, and their mean values were obtained. For inhomogeneous tumors and tumors with ringlike enhancement, representative well-enhanced peripheral areas and central areas with predominantly poor enhancement were selected on the basis of visual analysis of images. A curve for the ROI (tumor bed) in each examination was generated by computing the signal intensity in the ROI versus time. Pharmacokinetic analysis was based on a model described by Tofts and Kermode (13) that assumes that the contrast agent is distributed in the body in a multicompartmental exchange system (see Appendix).
Statistical Analysis
The differences between the two groups were tested for statistical significance by using an unpaired t test or 
2 test. A P value of less than .05 was considered to indicate a statistically significant difference. The parameters that may have affected treatment response were analyzed by using a multiregression analysis. All statistical analysis was performed by using statistical software (STATVIEW for Macintosh, version 4; Abacus, Berkeley, Calif).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONAPPENDIXREFERENCES |
|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
The pattern of enhancement or increased permeability in the region of the tumor correlated with the incidence of durability or local recurrence. Failure of radiation therapy or local recurrence as a function of the pharmacokinetic parameters and pattern of enhancement at pretherapeutic dynamic MR imaging is shown in Table 3. Tumors with a homogeneous enhancement pattern showed results with treatment that were significantly better than those of peripheral masses or those with predominantly poor enhancement (P = .01, 2 test). Tumors with increased permeability also had a higher incidence of good response to treatment (P < .01, unpaired t test). Regarding the extracellular fraction, good responders tended to have lower f values compared with those of poor responders, but this difference was not statistically significant.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONAPPENDIXREFERENCES |
|---|
A pharmacokinetic model has been used for the quantitative analysis of the kinetics of signal intensity enhancement that reflects the microcirculation of tumors (6,13). The two-compartment model describes the kinetics of the distribution of contrast medium in tissue and allows analysis of signal intensity–time curves, which yields values for the pharmacokinetic parameters. The parameter k is a tissue-specific transport parameter that describes the transfer of contrast medium between the plasma compartment and interstitial space. This parameter is mainly influenced by vascular permeability. The parameter f is the ratio of extracellular volume to intravascular volume, which contributes to the signal intensity. This parameter is influenced by the fraction of tumor vascular volume Vp and the fraction of extracellular leakage space Ve.
At pharmacokinetic analysis, it has been shown that vascularization (16), permeability, and interstitial space in cervical cancers are substantially greater than those of normal tissue; unlike the vascularity in normal cervical stroma, there will be more space in tumors to allow for the transportation and accumulation of contrast material (12,17). In vessels in benign tissues, the endothelium is surrounded by a basement membrane; this membrane may be damaged or missing in malignant tumors (17,18). On the basis of visual and pharmacokinetic analysis, we found regions that exhibited marked contrast enhancement due to the presence of dense and leaky capillaries in the periphery of cervical cancer tumors. Our evaluation also revealed that histopathologically proved zones of abundant tumor cells, mostly seen in the periphery of large tumors or in the entirety of most small tumors, had mean k and f values that were significantly higher and lower, respectively, than those of the center of large tumors.
It has been recently reported that dynamic MR imaging may play a role in the distinction between aggressive and nonaggressive tumors. Hawighorst et al (12) observed that the tumors associated with a more pronounced degree of tumor angiogenesis have a markedly shorter contrast medium exchange rate; that is, such tumors show more rapid first-pass contrast enhancement at dynamic MR imaging. Because tumor angiogenesis has been implicated as a predictor of tumor recurrence (8), fast or intense enhancement in early-phase dynamic MR imaging would correspond to more aggressive tumor behavior (6). However, this was not proved at clinical observation. Tumor angiogenesis determined by dynamic MR imaging findings did not correlate with tumor aggressiveness, as evaluated by the depth of tumor invasion and lymph nodal metastasis (5). This finding may imply that tumor angiogenesis is not a single predictor of tumor aggressiveness. Other factors such as tumor size or histologic grade may also be important.
Tumor vascularity and oxygenation status are two of the most important factors in determining the sensitivity of a tumor to irradiation or chemotherapy. Numerous investigations were performed to elucidate the role of blood supply and oxygenation, which may affect the response to radiation therapy in cervical cancers (19,20). To date, most methods for the clinical assessment of vascularity and tumor oxygenation frequently require the performance of invasive procedures, such as repeated oxygen-electrode placement (21,22). Direct measurements of tumor oxygenation by means of oxygen-electrode placement in cervical cancers also show a high correlation between radiation therapy failure and decreased oxygen levels (23).
Dynamic MR imaging findings support the concept that tumor blood supply, which probably relates to oxygenation, plays an important role in the prediction of the response to radiation therapy. A failure to respond to radiation therapy is often difficult to predict before therapy, when major decisions about the treatment regimen are made. Therefore, it is essential to identify the high-risk group of patients who are likely to be resistant to radiation therapy. If poor tumor response can be predicted before the initiation of radiation therapy, other more aggressive treatment modalities (the addition of chemotherapy, etc) should be considered.
Our results suggest that marked enhancement, which indicates high tumor perfusion or good blood supply, was associated with higher local control. In these patients, vascular supply was at an optimum and hypoxia was at a minimum before therapy, which led to better tumor control (9). With dynamic MR imaging techniques, we can distinguish poorly perfused areas that indicate hypoxic areas from well-perfused areas in vivo. In poorly perfused areas, fibrotic tissue was dominant on radiologic and histopathologic correlation.
There are a few limitations in this study. First, although we performed MR imaging and histopathologic correlation in the surgical group and although the results were applied to the radiation group, there was no evidence that the two groups were similar. However, because patient age, tumor grade, and tumor size were not significantly different, the results of radiation therapy may be extended to radiation therapy.
Second, for the analysis of tumor perfusion, we could have used a faster sequence for dynamic MR imaging at the expense of spatial resolution. However, a faster sequence is not required to analyze the dynamic curve of cervical cancers because the changes in signal intensity on dynamic MR images are gradual (4). Therefore, we chose to use sequences with moderate temporal resolution and excellent spatial resolution.
Third, ROIs were obtained in regions that were thought to be either poorly enhanced central areas or well-enhanced peripheral areas in inhomogeneous tumors or tumors with ringlike enhancement. This analysis has a potential sampling error. Although a pixel-by-pixel analysis would have been desirable, it requires the use of a special computer program that is not usually available. Instead, multiple ROIs were placed without the knowledge of clinical outcome of the patients, and their means were obtained.
Finally, because our observational period was relatively short because patients received various treatments for incomplete response or tumor recurrence, we could not analyze patient survival based on the parameters obtained at MR imaging. A further follow-up study with a larger number of patients may be required to confirm our results.
In conclusion, our results suggest that contrast-enhanced dynamic perfusion MR imaging studies performed with pharmacokinetic analysis before radiation therapy can be used to distinguish poorly perfused fractions from well-perfused fractions. Areas of increased contrast enhancement are mainly composed of abundant cancer cell fascicles, whereas areas of poorly perfused area are composed of fibrous tissue with scattered cancer cells. This method may offer important information about treatment outcome in patients with cervical cancer who are treated with radiation therapy; it can be helpful in the identification of patients who respond poorly to conventional radiation therapy.
|
APPENDIX |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONAPPENDIXREFERENCES |
|---|
Signal intensity measurements were made in ROIs that showed enhancement within the tumor on the dynamic MR image. The ROI includes an intratumoral plasma component (f1), an intratumoral intracellular component (f2), and an intratumoral interstitial component (f3). The triexponential concentration curve of gadopentetate dimeglumine was fitted to a theoretic model based on compartmental analysis. By using this method, the transfer constant (or permeability surface product per unit volume of component k) was measured, and the extracellular fraction f was measured as follows: f = f3/f1, where f1 represents tumor vascular volume, and f3 represents interstitial volume.
It is assumed that t = 0 when gadopentetate dimeglumine has uniformly penetrated the blood plasma but has not fully penetrated the interstitial space. The concentration of gadopentetate dimeglumine in the blood plasma of the ROI is assumed to be Cp(t), and its concentration in the interstitial space is assumed to be Ct(t). The following relationship can be obtained between Cp(t) and Ct(t):
Cp(t) can be expressed by the following equation:
If this model is used, Equation (A1) can be resolved as the following:
The concentrations in plasma Cp(t) in the ROI and interstitial space Ct(t) contribute to the measured values of signal intensities on MR images of tissue other than that of the central nervous system. Since values with T1, T2, and proton density are different from Cp(t) and Ct(t), it cannot be determined that Cp(t) and Ct(t) always contribute equally to the measured values. The concentration of gadopentetate dimeglumine contributes to the values of T1 and T2 as follows: 1/T1(t) = 1/T1(0) + R1C(t) and 1/T2(t) = 1/T2(0) + R2C(t), where R1 = 4.5S-1 mol · L-1 and R2 = 5.5S-1 mol · L-1.
The measured values are in proportion to the concentration of gadopentetate dimeglumine if the computation is made with the conditions where TR is much greater than T2 in the fast gradient-echo sequence. Thus, when portions that contribute to measured values of Cp(t) and Ct(t) are assumed to be f1 and f3, the measured values can be expressed by the following formulas and equation: intensity 
f1 Cp(t) + f3Ct(t), intensity Cp(t) + (f3/f1)Ct(t), and intensity = constant [Cp(t) + fCt(t)].
In this method, the time-concentration curve is assumed to be the ideal time-concentration curve of blood plasma at the time when 0.1 mmol/kg gadopentetate dimeglumine is intravenously and rapidly injected. This method was approximated in all examples (24), where A1 = 0.40 kg/L, A2 = 0.48 kg/L, m1 = 0.0023 sec-1, and m2 = 0.00018 sec-1. If these values are applied to Equation (A2), Cp(t) = 0.40exp(-0.0023t) + 0.48exp(-0.00018t).
If Equation (A3) is substituted for intensity = constant [Cp(t) + fCt(t)],
In this model, k corresponds to the speed with which gadopentetate dimeglumine spreads from blood plasma to the interstitial space at different concentrations, and the rate of proportion with which Cp(t) and Ct(t) contribute to the measured values is a parameter is obtained.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, Y.Y.; study concepts, T.B.; study design, Y.B.; definition of intellectual content, Y.B.; literature research, S.I.; clinical studies, T.B.; data acquisition, R.N.; data analysis, H. Ohtake; statistical analysis, Y.Y.; manuscript preparation, Y.Y.; manuscript editing, R.N.; manuscript review, M.T., H. Okamura.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONAPPENDIXREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  H. Hricak, C. Gatsonis, F. V. Coakley, B. Snyder, C. Reinhold, L. H. Schwartz, P. J. Woodward, H. K. Pannu, M. Amendola, and D. G. Mitchell Early Invasive Cervical Cancer: CT and MR Imaging in Preoperative Evaluation ACRIN/GOG Comparative Study of Diagnostic Performance and Interobserver Variability Radiology, November 1, 2007; 245(2): 491 - 498. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. J. Choi, J. K. Kim, N. Kim, K. W. Kim, E. K. Choi, and K.-S. Cho Functional MR Imaging of Prostate Cancer RadioGraphics, January 1, 2007; 27(1): 63 - 75. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C Lopez and S Chakravarti Imaging of cervical cancer Imaging, August 1, 2006; 18(1): 10 - 19. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. H. Dennis Cheong, C. C. Tchoyoson Lim, and T. S. Koh Dynamic Contrast-enhanced CT of Intracranial Meningioma: Comparison of Distributed and Compartmental Tracer Kinetic Models--Initial Results Radiology, September 1, 2004; 232(3): 921 - 930. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A R Padhani MRI for assessing antivascular cancer treatments Br. J. Radiol., December 1, 2003; 76(suppl_1): S60 - S80. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Kaur, P. M. Silverman, R. B. Iyer, C. F. Verschraegen, P. J. Eifel, and C. Charnsangavej Diagnosis, Staging, and Surveillance of Cervical Carcinoma Am. J. Roentgenol., June 1, 2003; 180(6): 1621 - 1631. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
