Dynamic computed tomography and color Doppler ultrasound of renal parenchymal neoplasms: correlations with histopathological findings.

BACKGROUND
We evaluated whether color Doppler ultrasound (US) had diagnostic accuracy equal to dynamic computed tomography (CT) and whether performing dynamic CT and Doppler US together would be more informative in preoperative diagnosis of renal solid tumors.


METHODS
A total of 110 renal solid tumors smaller than 7 cm were evaluated with dynamic CT and Doppler US. We compared the enhancement and the color flow patterns with the histopathological subtypes.


RESULTS
Eighty-seven (95.6%) of 91 clear cell carcinomas showed rich enhancement in the cortical nephrographic phase (CNP) and 82 (90.1%) of them had color flow in the Doppler US. Of the total of 110 tumors, nine (8.1%) did not show color flow in spite of rich enhancement in the CNP. Conversely, eight (7.2%) of the 110 tumors showed color flow in spite of poor enhancement, including two chromophobe cell carcinomas and two metastatic renal tumors.


CONCLUSIONS
The enhancement pattern in dynamic CT and the color flow pattern in Doppler US were different among the subtypes of RCC. Color Doppler US had diagnostic accuracy equal to dynamic CT in most patients with renal solid tumors. Although Doppler US may play a unique role in the diagnosis of some renal parenchymal solid tumors, it is sufficient to perform dynamic CT alone for diagnosis of clear cell carcinoma.


INTRODUCTION
In recent years, computed tomography (CT) and ultrasound (US) have dramatically improved the early detection of renal masses. Most renal cell carcinomas (RCCs) are characterized by abnormal vascular structures (1). Dynamic CT is an established method of imaging renal masses and evaluating their vascularity via pooling of contrast medium in the tumor (2). Clear cell carcinoma (conventional RCC) is described as an attenuated tumor in the arterial phase of dynamic CT (3). The enhancement pattern in dynamic CT has been reported to be different among the subtypes of renal cell carcinoma (4). To the best of our knowledge, however, there has been no report showing the patterns of color Doppler US among the subtypes of renal neoplasms. Moreover, some patients are allergic to the contrast medium used for dynamic CT and patients are also exposed to a large amount of X-rays in CT scanning. Hence there is a need for diagnosis of RCC by another method less invasive than CT. Doppler US allows for non-invasive assessment of vascular flow signals from neovascularity in tumors (5). Thus, we first hypothesized that color Doppler US may also show different patterns like dynamic CT findings. Second, we also hypothesized that we could obtain more accurate information about the histopathological subtypes of renal parenchymal neoplasms if both dynamic CT and color Doppler US were performed. In this study, we compared the histopathological distribution of the dynamic CT findings for renal solid tumors with the color Doppler US findings and evaluated whether dynamic CT is more informative in preoperative diagnosis of renal solid tumors.

PATIENTS
Between August 1996 and May 2001, 110 patients with solid masses smaller than 7 cm in diameter and with pathological results confirmed by surgical removal of tumors were eligible for this study. The patients (75 men and 35 women) ranged in age from 22 to 85 years (median 62 years) and tumor size ranged from 1.0 to 7.0 cm (median 3.6 cm) in diameter. These patients were suspected to have RCC by US and/or CT at other hospitals and were introduced to us for further evaluation and treatment of those tumors. By initially performed B-mode US, angiomyolipomas (AMLs) were excluded.

DYNAMIC CT
Dynamic CT was performed using a Toshiba X-Vigor with 7 mm/s table feed for 7 mm slice helical scans. An unenhanced scan was carried out initially to obtain baseline attenuation values of lesions and to identify calcification. After bolus administration of 100-120 ml of contrast medium intravenously (injection rate 2.5-3.0 ml/s), two phases of renal enhancement were recognized. Initially, 30 s after the contrast medium injection, there was enhancement of the cortex, but not the medulla, so that cortical nephrographic differentiation was seen [cortical nephrographic phase (CNP)]. Five minutes after the injection, additional scanning was performed to obtain images of its excretion by the pelvic caliceal system (excretory phase). If the tumor density in the CNP was higher than that of the renal medulla, it was defined as rich enhancement. If it was lower than that of the renal medulla but higher than that in the preenhancement phase, it was defined as poor enhancement.

COLOR DOPPLER US
Color Doppler US examinations were performed with commercially available real-time scanners (Toshiba SSA380A) by the same senior radiologist (Y.M.). The radiologist had no information about the dynamic CT findings. Patients were examined in the supine and lateral decubitus positions, using transverse, intercostal and parasaggital scanning. The insonating frequency of the sector scanner was 4.7 mHz. The Toshiba unit incorporates a real-time scanner and range-gated pulse Doppler velocity meter. The wall filter was set as low as possible and the real-time B-mode was used to locate the tumors. The Doppler US scanning was performed under conditions of no color flow in the normal renal parenchyma. If color flow was detected in the tumor, it was defined as positive flow.

CLASSIFICATION
Histopathological findings were reviewed for the subtypes of neoplasms, according to the classifications of the Union Internationale Contre le Cancer (UICC) and the American Joint Committee on Cancer (AJCC) (6). By the findings and estimation of detection of early enhancement patterns in dynamic CT and color flow in Doppler US, four distinct patterns could be identified: group 1, tumors with both rich enhancement and color flow; group 2, with rich enhancement and without color flow; group 3, with poor enhancement and with color flow; and group 4, with poor enhancement and without color flow.

HISTOPATHOLOGICAL DISTRIBUTION OF ALL 110 TUMORS
Histopathological examination revealed that, of the 110 neoplasms, 91 were clear cell carcinomas, five were papillary RCCs, four were granular cell carcinomas, three were benign renal tumors, two were chromophobe cell carcinomas, two were spindle cell carcinomas, two were metastatic renal tumors and one was a collecting-duct carcinoma.
One of the metastatic tumors was malignant melanoma. There were no other metastatic sites in the patient. The other was thyroid papillary cancer. Although the patients also had lung, retroperitoneal and mediastinal lymph node metastases, primary renal neoplasm could not be denied because only the renal tumor had grown in a short time.
The three benign tumors were composed of an oncocytoma, a leiomyoma and an AML. (TABLE 1)

GROUP 1 (n = 87)
Eighty-one (93.1%) of the 87 tumors were clear cell carcinomas. The remainder consisted of one granular cell carcinoma, one papillary RCC, one spindle cell carcinoma and three benign tumors (an AML, an oncocytoma and a leiomyoma). All the benign tumors were diagnosed as renal cell carcinoma before the surgical treatment.

GROUP 2 (n = 9)
Six (66.7%) of the nine tumors were clear cell carcinomas. The others were two granular cell carcinomas and one collectingduct carcinoma.

Comparison of Doppler US and dynamic CT GROUP 3 (n = 8)
The group three tumors consisted of two chromophobe cell carcinomas, two metastatic tumors, one clear cell carcinoma, one granular cell carcinoma, one papillary RCC and one spindle cell carcinoma. The histological diagnoses of the metastases were papillary thyroid cancer and malignant melanoma.

GROUP 4 (n = 6)
Three papillary RCCs and three clear cell carcinomas were identified. Two of the clear cell carcinomas had atypical structures; they were only tubular and solid with a small tubular part, respectively. The remaining one had severe arteriosclerosis of the renal artery.

DISCUSSION
Recent advances in US, CT and magnetic resonance imaging (MRI) techniques have enabled us to detect incidentally renal cell carcinomas (7,8). In our institute, most renal parenchymal neoplasms were first suspected by physicians to be RCCs in the process of examinations for non-urological diseases. After the diagnosis as a renal parenchymal solid tumor other than AML, we performed dynamic CT and confirmed the diagnosis as RCC before the surgical treatment because dynamic CT has been the most readily available method for diagnosis of RCC, including subtypes (3,4). However, it is essential for a diagnosis of RCC to use contrast medium in dynamic CT and patients are irradiated with a large amount of X-rays. Moreover, we sometimes cannot confirm the diagnosis of a renal malignant tumor by dynamic CT even if AML is denied by B-mode US. Therefore, we studied the efficacy of color Doppler US for the diagnosis of renal parenchymal neoplasms. Doppler US is at least as accurate as CT in staging of RCC (9) and may improve the accuracy of US determination of malignancy (10). However, to the best of our knowledge, there has been no report on correlations between the color flow patterns and the subtypes of RCC. The relationship between dynamic CT and color Doppler US findings also has not been reported. We previously reported that clear cell carcinoma with alveolar architecture showed a highly attenuated area in the CNP of dynamic CT (3). Jinzaki et al. (4) also reported that clear cell carcinoma showed a peak attenuation value in the CNP of >100 HU, whereas for other subtypes the values were <100 HU. The attenuation patterns of the tumors in this study were mostly consistent with those reported previously. Since the color flow positive rate of our tumors was similar to the dynamic CT positive rate, it was suggested that color Doppler US was as readily applicable as dynamic CT for diagnosis of renal parenchymal neoplasms. However, there were some tumors without color flow in Doppler US in spite of rich enhancement in dynamic CT. Conversely, there were also some tumors with color flow in spite of poor enhancement in CT.
In our series, color Doppler US showed color flow in chromophobe cell carcinomas despite the fact that dynamic CT showed poor enhancement of these tumors. Although it is too early to discuss our small number of chromophobe cell carcinomas, it has been reported that chromophobe cell carcinoma has a peak attenuation value in the CNP in dynamic CT of <100 HU (4). It may be better to perform additional color Doppler US if dynamic CT does not demonstrate a highly attenuated tumor.
Doppler US also showed color flow in metastatic renal tumors. However, the number of our patients was too small to analyze the characteristics of such tumors. The findings of dynamic CT and color Doppler US might be different for each primary tumor.
In this study, three benign tumors were diagnosed as RCC both by dynamic CT and color Doppler US. Jinzaki et al. (4) reported that it was too difficult to differentiate RCC and other benign tumors (oncocytoma and metanephric adenoma) by dynamic CT. We suggest that there is no difference between the false-positive rates of dynamic CT and of Doppler US in diagnosis of renal parenchymal tumors and that another diagnostic method is necessary to differentiate between them. In contrast, there were six tumors (three papillary RCCs and three clear cell carcinomas) diagnosed as non-RCC by both CT and Doppler US. Most papillary RCCs show hypovascularity (11) and lower enhancement in the cortical nephrographic phase of dynamic CT than clear cell carcinoma (4). Choyke et al. (12) reported that the tumors of patients with hereditary papillary renal cancer syndrome posed some diagnostic difficulties because they could be missed by US, were small and enhanced poorly on CT. Moreover, even if the histological cell type is clear cell carcinoma, the hypervascularity on the CNP of dynamic CT is not shown if the architecture of the tumor is not the alveolar type (3). Therefore, new diagnostic methods are needed for the diagnosis of those tumors. It may be possible to clarify the discrepancies between the Doppler US and the dynamic CT findings by using some new diagnosis modalities, e.g. contrast-enhanced Doppler US. However, a prospective study with a large number of patients is needed to clarify the efficacy.
The reproducibility of color Doppler US might be doubtful, although a senior radiologist performed color Doppler US for all the patients in our series. Dynamic CT is superior to Doppler US in this respect. However, color Doppler US is performed safely for patients who are allergic to contrast medium or who are pregnant. Although it is sufficient to perform dynamic CT alone for the diagnosis of renal solid tumors in most patients, color Doppler US can be used instead of dynamic CT in patients whose tumor is poorly attenuated or who have problems with using contrast medium, exposure to radiation, etc.
In conclusion, we can diagnose renal solid tumors by dynamic CT alone in most patients, although the enhancement pattern in dynamic CT and the color flow pattern in Doppler US are different among the subtypes of RCC. Doppler US may play a unique role in the diagnosis of some renal parenchymal solid tumors. However, more data on chromophobe cell carcinoma, metastatic renal cancer, etc., are needed.