Alveolar Architecture of Clear Cell Renal Carcinomas «5.0 cm) Show High Attenuation on Dynamic CT Scanning

Transcription

1 Jpn J Clin OncoI1999;29(4) Alveolar Architecture of Clear Cell Renal Carcinomas «5.0 cm) Show High Attenuation on Dynamic CT Scanning Hiroyuki Fujimoto1, Fumihiko Wakao 2, Noriyuki Moriyama 2, Kenichi Tobisu1, Michiie Sakamoto 3 and Tadao Kakizoe 1 Departments of 1Urology and 2Diagnostic Radiology, National Cancer Center Hospital and 30ivision of Pathology, National Cancer Center Research Institute, Tokyo, Japan INTRODUCTION Background: Toestablish the correlation between tumor appearance on CT and tumor histology in renal cell carcinomas. Methods: The density and attenuation patterns of 96 renal cell carcinomas, each :::;5 cm in greatest diameter, were studied by non-enhanced CT and early and late after bolus injection of contrast medium using dynamic CT. The density and attenuation patterns and pathological maps of each tumor were individually correlated. Results: High attenuated areaswere presentin 72 of the 96 tumors on early enhanced dynamic CT scanning. All 72 high attenuated areas were of the clear cell renal cell carcinoma and had alveolar architecture. The remaining 24 tumors that did not demonstrate high attenuated foci on early enhanced scanning included three clear cell, nine granular cell, six papillary, five chromophobe and onecollecting ducttype. Withrespect to tumorarchitecture, allclearcelltumors of alveolar architecture demonstrated high attenuation on early enhanced scanning. Conclusion: Clear cell renal cell carcinomas of alveolar architecture show high attenuation on early enhanced dynamic CT scanning. A largernumber of patients are indispensable to obtaining clear results. However, these findings seem to be an important clue to the diagnosis of renal cell carcinomas as having an alveolar structure. Key words: renal cell carcinoma - computed tomography - attenuation Recently, the wide use of computed tomography (CT) and trans-abdominal ultrasound have allowed earlier detection of small renal tumors 0-3). A common radiological characteristic of renal cell carcinomas is increased vascularity. Selective renal arteriography has been performed to estimate tumor vascularity as the final and definitive diagnostic step in the evaluation oflarge renal masses (4). However, the histological architectures representing radiological features such as arteriovenous fistulae and pooling of contrast media have not been definitively determined. Dynamic thin-section CT with administration of contrast medium by bolus peripheral injection has been reported to be superior to conventional enhanced CT in evaluating tumor vascularity (5,6). If a correlationbetweentumor vascularity and tumor morphology can be established, a more precise diagnosis of renal cell carcinoma and different diagnosis between renal cell carcinoma Received November 5, 1998; accepted December 28, 1998 For reprints and all correspondence: Hiroyuki Fujimoto, Division of Urology, National Cancer Center Hospital, 1-1, Tsukiji 5 chome, Chuo-ku, Tokyo,I ,Japan from other renal masses including oncocytoma, angiomyolipoma with low content of lipid structure, complicated cyst and benign mesenchymal tumor may be achieved preoperatively. In an attempt to evaluate this correlation, the density and attenuation patterns of 96 renal cell carcinomas were studied. MATERIALS AND METHODS PATIENTS A total of 96 renal cell carcinomas with individual diameters of :::;5 ern from 96 consecutive patients with renal cell carcinomas were evaluated between January 1990 and April The larger renal cell carcinomas tend to have the more heterogeneous architecture, thus leading to 'misunderstanding of the accurate correlation between tumor mapping and CT findings. For the first steps of this analysis, renal cell carcinomas :::;5 em in greatest diameter were selected. Patients were subjected to dynamic CT scanning followed by radical or partial nephrectomy. All specimens were examined by step-section. The size of the tumors ranged from 1.3 to 5 cm (mean 3.5 ern). Tumor size was determined using the greatest diameter of the freshly excised specimens Foundation for Promotion of Cancer Research

2 Jpn J Clin OncoI1999;29(4) 199 D iso. low Diso Dhigh.markedly -low D low Figure 1. IT density patterns in plain (A) and attenuation pattern in early enhanced (B) and late enhanced (C) scans were classified as markedly low,low, iso or high in comparison with the renal medulla. The lower half is a schematic illustration. cr TECHNIQUE The initial scan was performed without contrast medium (plain scan). To obtain the early enhanced scan, scanning was begun s after machine-injected bolus administration of ml of contrast medium (injection rate ml/s). The equipment used was a Toshiba X-Vigor with 7 rum/second table feed for 7 mm slice helical (spiral) scans (6). Additional scanning was carried out to 5-7 min after initial injection of contrast medium to record late enhancement (late enhanced scan). Tumor density in plain scan and tumor attenuation in early and late scan were classified as markedly low, low, iso and high in comparison with that of renal medulla (Fig. 1). PATHOLOGICAL ANALYSI S Horizontal step-sectioning of excised kidne y at - 1 ernwidth was adopted to compare the pathological finding s with CT findings. Macro scopic views of the greatest diameters were recorded photographi cally. Histopathological whole mapping of the greatest diameters was carried out in accordance with the Arme d Forces Institute of Pathology (APIP) classification of renal cell carcinoma (7). In this classification, cell types are listed as clear cell, granular cell (chromophilic), papillary, chromophobe cell, sarcomatoid and collecting duct and morph ological architecture as alveolar, acinar and cystic in clear cell renal cell carcinom a. The tumor grade was determined by Fuhrman's grading system (8). The 96 tumors were classified as follows: 75 clear cell, nine granular cell, six papillary, five chromophobe cell and one collecting duct type. No sarcomatoid tumors were present in this series. A NALYSIS OF C ORRELATION BETWEEN T UMOR M ORPHOLOGY AND CT FINDINGS If the tumor showed heterogeneous attenuation on early scan, three isolated high, iso or low attenuation areas on early enhanced scanning of that tumor 's greatest diameter were exc ised and three areas (about 2-3 mm 2 ) were selected for analysis. If the tumor showed homogeneous attenuation, 1-3 random areas were selected to assess the histological difference. The density pattern s on plain scanning and the attenuation pattern on late enhanced scanning of these same areas were examined. Following whole mapping of each tumor, tumor morphology and the CT finding s were correlated. Because a single tumor had up to three different evaluable areas on CT scanning, a total of 189 areas from 96 tumors were analyzed in this series. Although the cell type of a given tumor was uniform, the cell architecture within a single tumor varied in accordance with the CT fmdings described below. RESULTS In the early enhanced scans, 72 (75%) of96 tumors demonstrated at least one high attenuated area. The remaining 24 (25%) tumors did not demonstrate any areas of high attenuation. All 72 high attenuated tumors were clear cell rumors. The remaining 24 tumors included three clear cell, nine granular cell, six papillary, five chromophobe and one collecting duct type. Out of a total of 75 clear cell tumor s observed in this series, only three did not show high attenuated areas on the early enhanced scanning. Of these three clear cell tumors, two had an acinar structure without alveolar architecture and the remaining tumor had a solid architecture. CT findings in the early enhanced scans showed a good correlation with gross findings and pathological mapping of the

3 200 Dynamic CT ofrenal cell carcinoma normal kidney D alveolar.grade 1 D cystic.grade 1 Figure 2. Correlation betweendensity patternon dynamicct in plain(a)and attenuation pattern inearly enhanced (B)and lateenhanced (C) scan, macroscopic view (D) and pathological mapping (E). o alveola r, grade 1 18 alveo-acinar, grade 2 o acinar, grade 1 Figure 3. Clearcell renalcell carcinoma,the attenuation patternon enhancedctand tumormorphologyexhibiteda goodcorrelation, Alveolarstructure demonstrates high attenuation aridcystic lesions demonstratelow attenuation in early enhancedscans. high attenuated tumors (Figs 2 and 3). The 189 identified areas included clear cell tumors of the following architectures: 100 alveolar, 35 acinar and 20 cystic. The remaining 34 areas included six papillary cell tumors of a papillary and one of a papilla-acinar structure, seven granular cell tumors of alveolar structure, five chromophobe cell ' tumors of broad alveoli architecture, one collecting duct tumor of dilated tubule architecture (7) and 14 fibrotic or necrotic lesions.

4 Jpn J Clin OncoI1999;29(4) 201 Table 1. Correlation of non-enhanced CT density with histology and tumor density Cell type Clear Papillary Granular Chromophobe Collecting duct Tumor architecture CT density Low Iso High Alveolar 23/100 (23.00%) 77/100 (77.0%) 0/100 Acinar 20/35 (57.1%) 14/35 (40.0%) 1/35 (2.9%) Cystic 14/20 (70.0%) 6/20 (30%) 0/20 Papillary 0/7 1/7 (14.2%) 6/7 (85.7%) Alveolar arrangement 0/6 6/6 (100%) 0/6 Broad alveoli 0/5 5/5 (100%) 0/5 Dilated tublus 1/1 (100%) 0/1 0/1 On unenhanced scanning, 85.7% papillary renal ceii carcinoma demonstrated high density. Almost all other tumors showed iso or low density. Table 2. Correlation of tumor attenuation with histology Cell type Clear Papillary Granular Chromophobe Collecting duct Tumor architecture Attenuation in early-late scan on enhanced dynamic CT Low-m.low* Low-low Iso-m.low* Iso-low High-low Alveolar /100 (3/0%) 97/100 (97%) Acinar a /35 (l00%) 0 Cystic 16/20 (80.0%) 4/20 (20%) Papillary a a 0 7/7 (100%) 0 Alveolar arrangement 0 1/6 (100%) 0 5/6 (83.3%) 0 Broad alveoli /5 (100%) 0 Dilated tubulus a 0 1/1 (100%) a 0 Only clear cell tumors of alveolar architecturedemonstrated high attenuationon early enhanced scanning. Granular, chromophobe, papillary and coliecting duct types of renal cell carcinomas demonstrated iso-attenuation in early enhanced scans except for one granulartumor with low attenuation. Fibrosis and necrosis are excluded from this table. *Markedly low. The density patterns in plain scan and the attenuation patterns in early and late scan of these 175 areas (all except the 14 fibrotic or necrotic lesions) are shown in Tables 1 and 2. On plain scans, 6n (85.7%) papillary and one acinar clear cell tumors revealed high density. Almost clear cells showed low or iso-density on plain scan. On early enhanced scanning, only clear cell tumors of alveolar structure demonstrated high attenuation. Ninety seven (97.0%) of 100 lesions with alveolar structure revealed high attenuation in early enhanced scans and low attenuation in late enhanced scans. All acinar structure tumors showed iso-attenuation in early enhanced scans and low attenuation in late enhanced scans. Sixteen (80%) of 20 cystic tumors exhibited low attenuation in early enhanced scans and markedly low attenuation in late enhanced scans. Papillary (Fig. 4), granular (Fig. 5) and chromophobe (Fig. 6) types demonstrated iso-attenuationin early scan. There were no high attenuated cell types except clear cell types. There was no definite correlation between tumor grade and tumor attenuation. DISCUSSION Most renal cell carcinomas are well known to be hypervascular tumors on angiography (4). However, an accurate correlation between tumor vascularity and pathological architecture has not been established. In analyzing 96 renal cell carcinomas ~5 cm in diameter, it was found that the attenuation pattern on early enhanced dynamic CT scanning correlated well with tumor cell type and pathological architecture. Only clear cell tumors with alveolar architecture demonstrated high attenuation compared to that of renal medulla on early enhanced scanning. Tumors with acinar architecture had almost the same attenuation as that of renal tubules. Cystic tumors showed low attenuation in early enhanced scans and markedly low attenuation in late enhanced scans. These results led us to conclude that the hypervascularity of most renal cell carcinomas on early dynamic CT scanning is due to the alveolar architecture of clear cell tumors in small (~5.0 em ) renal cell carcinoma. A few clear cell tumors did not included areas of alveolar architecture and that is why these tumors did not demonstrate high attenuated lesion. For iso-attenuated tumors, it is difficult to reach an accurate diagnosis from the attenuation pattern of dynamic CT. Iso-attenuated tumors included papillary, chromophobe, granular, collecting duct and rare clear cell types. In our series, six of seven papillary tumors were iso-attenuation tumors on early enhanced scanning and showed high density on plain CT. High density on non-enhanced CT scanning may be one of the characteristics of papillary renal cell carcinomas. No other features of these tumors were established.

5 202 Dynamic CT ofrenal cell carcinoma D papillary, grade 2 Ii] papillo-acinar, grade 2 Figure 4. Papillary renal cell carcinoma. In plain scan the tumor showed high density and no high attenuation in early scan. D alveolar arrangement, G3 Figure S. Granular renal cell carcinoma. In early scan the tumor was homogeneous and showed no high attenuation. Analysis of CT number may be useful in theses iso-density tumors. Fairly recently, Thoenes et al. (9) reported a new classification system for renal cell carcinoma including chromophobe cell tumors. According to their classification based on the histogenesis of renal cell carcinomas, the clear cell type is derived from renal proximal tubules whereas the chromophobe cell type comes from distal tubules. In early dynamic CT scanning, clear cell tumors of alveolar architecture demonstrated a high attenuation that was almost equal to the attenuation of renal cortex. In contrast, the chromophobe cell type showed iso-attenuation with the renal medulla and had lower attenuation than the renal cortex. These phenomena may explain the good correspondence with tumor vascularity and histogenesis reported by Thoenes et al. Recent cytogenetic studies have revealed that papillary and chromophobe cell tumors are special types of renal cell carcino-

6 Jpn J cu«oncol1999;29(4 ) 203 normal parenchyma D broad alveoli Figure 6. Chromophobe renal cell carcinoma. In the plain scan, the tumor showed iso-density and in early scan iso-auenuation. rna ( 10-12). Clinically, our data suggest that the vascularity of clear cell renal carcinomas differs from that of non-clear cell renal carcinom as. Considering these density and attenuation patterns,.clear cell tumors appear to be a different entity from other variants of renal cell carcinoma such as papillary, chromophobe, collecting duct and granular cell types. Obviously, further histogenetic or cytogenetic studies will be necessary to elucidate the clinicopathological significance of these findings and whether they result in different behavior among these types of renal cell carcinoma. In conclusion, clear cell renal cell carcinoma of alveolar architecture expressed high' attenuation on early enhanced dynamic CT scanning, whereas those of acinar architecture showed iso-attenuation with the renal medulla. The papillary, granular, collecting duct and chromophobe cell types showed iso-attenuation on early enhanced dynami c CT scanning. Expression of high attenuation on early enhanced dynamic CT scanning seems to be an important clue that a renal cell carcinoma is a clear cell tumor of alveolar architecture. Acknowledgment This study was supported by the second term Comprehensive 10 Year Strategy for Cancer Control of the Health and Welfare Ministry of Japan. References I. Bosniak MA. The small (:'>3.0 em) renal parenchymal tumor. detection, diagnosis and controversies. Radiology 1991;179: Silverman SG, Lee BY, Seltzer SE, Bloom DA, Corless CL, Adams DF. Small «3 em) renal masses: correlation of spiral cr features and pathological findings. Am J RoentgenoI1994;1 63: Yamashita Y, Takahashi M, Watanabe 0, Yoshimatsu S, Ueno S, Ishimaru S, et al. Small renal cell carcinoma: pathologic and radiologic COrrelation. Radiol ogy 1992;184: dekemion JB, Belldegrun A. Renal rumors. In: Walsh PC, Retik AB, Stamey TA, Vaugh ED Jr, editors Campbell's Urology, 6th ed. Philadelphia: Saunders 1992; Foley WD, Berland LL, Lawson Tl.Srnith DF, Thorsen MK. Contrast enhancement technique for dynamic hepatic computed tomographic scan-' ning. Radiology 1983;147: Zeman RK, Fox SH, Silverman PM, Davros W, Carter LM, Griego D, et al. Helical (spiral) CT of the abdomen. Am.f Roentgenol 1993;160: Murphy WM, Beckwith JB, Farrow GM. Tumors of the kideny. In:Tumors of the Kidney, Bladder and Related Urinary Structures, Third Series, Fascicle 11. Washington, DC: Armed Forces Institute of Pathology 1993; Fuhrman SA, Lasky LC, Limas C. Prognostic significance of morphologic parameters in renal cell carcinoma. Am J Surg ParhoI 1982;6: Thoenes W, Starkel St, Rumpelt HJ, Moll R. Cytomorphological typing of renal cell carcinoma. A new approach. Eur UroI 1990;I8(suppI 2): Akhtar M, Kardar H, Linjawi T, McClintock J, Ali MA. Chromophobe cell carcinoma of the kidney. A clinicopathologic study of 21 cases. Am J Surg Pathol 1995; 19: II. Thrash-Bingham CA, Salazar H, Freed Il, Greenberg RE, Tartof KD. Genomic alterations and instabilities in renal cell carcinomas and their relationship to tumor pathology. Cancer Res 1995;55: Renshaw AA, Corless CL. Papillary renal cell carcinoma. Histology and immunohistochemistry. Am J Surg Parhol 1995;19:842-9.