In 1976, Mancilla Jimenez et al. first reported that certain papillary renal cell carcinomas originate from the collecting duct. Fleming and Lewi described six cases of CDC and presented diagnostic criteria to recognize it as a unique pathological subtype of RCC. CDC was listed as one of the major subtypes of RCC in both the 2002 and 2006 WHO classifications. The tumor interstitial inflammatory fibrosis and collagen secretion are obvious, and the tumor tissue is dense. The results of the pathological examination of the patients in the present study are consistent with the literature (Figure 8A-D).
The patients in this study showed a wide age range, with a mean age at diagnosis of 64 years and a 3:1 male predominance. Although these findings are consistent with those of previous studies[8,9], the demographic proﬁle also applies to RCC in general and is therefore not a useful discriminator. The clinical manifestations, which were not specific to CDC and can be found in other RCCs, included gross hematuria, waist and abdomen pain, abdominal fullness, and sometimes a palpable mass. Gross hematuria was observed in 11 (84.6%) patients, and six (46.2%) had low back pain or lumbar discomfort. One patient who had no clinical manifestations of urinary disease and presented with a low-grade fever and night sweats was admitted to the hospital for suspected tuberculosis.
Site and border
The pathological features of CDC are as follows: the initial site of the tumor is the renal medulla, which is grayish white or light yellow. The renal interstitial cells function as a scaffold for tumor cells to diffuse and infiltrate along the collecting duct to the renal pelvis and renal cortex. Therefore, most of the tumors occur in the medulla and infiltrate into the cortex and renal pelvis. The borders are irregular and most kidney tumors have an outward expansive growth pattern from the center. The normal renal parenchyma is displaced and bulging of the kidney contour may be present, as well as the formation of a pseudocapsule[17,19]. Young et al. reported that sarcomatoid RCCs and CDCs are more likely to have an irregular contour and an infiltrative pattern than other RCC subtypes. When used to discriminate sarcomatoid RCC and collecting duct carcinoma from other solid renal masses, an infiltrative spread pattern has a specificity of 93% (287/308) and sensitivity of 82% (9/11), whereas an irregular contour has a specificity of 98% (303/308) and sensitivity of 64% (7/11).
CDC can be divided into medullary type, cortex-medullary type and cortex-medullary-pelvis type according to the infiltration site. The present cohort did not include patients with the simple medullary type. The masses were mostly of the cortex-medullary type (six cases, 46.2%) and the cortex-medullary-pelvis type (seven cases, 53.8%). This feature is different from the common renal clear cell carcinoma derived from cortical renal tubules. The common renal clear cell carcinoma is centered on the renal cortex and invades the medulla. The CDC tumor can rapidly grow toward the renal pelvis and cortex and destroy the renal pelvis and renal hilum structure, and may involve the upper ureter. Thus, assessing for the presence of an infiltrative spread pattern and an irregular contour can provide a simple, noninvasive means of discriminating CDC from other solid renal masses with a relatively high specificity, sensitivity, and negative predictive value.
Morphology and size
The morphology of tumors is closely related to the biological behavior and growth pattern of tumors. Despite its medullary derivation, almost all tumors exhibit focal cortical extension, and perinephric extension is also common. CDC tumors spread along the collecting duct during invasive growth, and there is fibrous tissue hyperplasia in the tumor stroma[5,8,28]. The tumors show a diffuse enlargement that follows the kidney contour or a certain kidney segment without clear boundaries, capsule, or pseudocapsule. This characteristic was observed in 12 of the 13 cases (92.3%) in our group. In addition, two patients showed a mass-type tumor. In one case, the inflammatory fibrous tissue proliferation was not significant. The tumor outline was bound by the tumor cell aggregation area. Therefore, we believe that the morphology of CDC tumors is related to the degree of tumor interstitial fibrous tissue hyperplasia. The extent of interstitial fibrosis is negatively correlated with the tumor outline. The morphological characteristics of the tumor are helpful to distinguish it from other kidney tumors.
The diameter of the tumors ranged from 5.12 to 12.40 cm (mean diameter, 8.48 ± 2.48 cm), and four cases had a diameter > 10 cm. In general, CDCs are relatively large. In large tumors, the medullary origin can be difﬁcult to determine. Fukuya et al. described the CT findings of small tumors measuring 3–4.5 cm. These lesions were all centered in the renal medulla; four of five protruded into the central sinus, and none showed exophytic growth; the reniform contour was preserved in all cases.
CT density and MRI signal
The originating organ of CDC is water-rich kidney tissue. The tumor stroma is characterized by increased fibrous tissue hyperplasia and collagenation. The density of the tumor parenchyma is higher than that of the surrounding normal tissues, which is a feature of this group of CDC tumors in non-enhanced CT. These characteristics differ from those of RCC arising from the renal cortex. The interstitial tissue is dense or collagen secretion is increased. Inflammatory fibroblastic tissue hyperplasia and abundant fibrous tissue components are characteristic of CDC tumors. These features are important for the pathological diagnosis of CDC and appear as a low signal on magnetic resonance T2WI.
The MR examinations reported by Pickhardt et al. (four cases) showed that the parenchymal components of all four tumors showed equal signals on T1WI. One tumor with multiple cystic components showed a low to high T1 signal for each cyst (including water, fat, bleeding, and other signals). On T2WI, the parenchymal components of the four tumors were lower than those of the normal renal parenchyma. There absence of a low signal at the edge of tumor indicated the presence of a pseudo-envelope observed on MRI. In the present study, MRI detected the tumor parenchyma as a low signal on T2W1 in one case. The area of cystic necrosis was detected as a high and low mixed signal on T2WI, with unclear separation from the surrounding normal renal parenchyma. There was no obvious ring-shaped low T2WI signal suggesting a pseudo-envelope.
The calcification in CDC tumors is probably due to the increased fibrous tissue in the tumor stroma, and calcium salts easily deposit in fibrous tissue. However, calcification was only present in two (15.4%) cases in this group. In contrast to other RCCs, calcification was also observed in only one case in the study by Seong et al.. Compared with the normal renal parenchyma, which is rich in water, the tumor tissue is denser and the interstitium shows inflammatory fibrous tissue hyperplasia. The tumor parenchyma is similar, showing a high signal on T1WI and a low signal on T2WI. Kato et al. described the signal intensity of CDC on T2WI as isointense or hypointense, which was thought to be due to hemosiderin deposition. Larger clear cell renal cell carcinomas tend to have a heterogeneous hyperintense signal on T2-WI, differentiating it from CDC.
Some liquid components are detected as low-density areas in tumors. The shape is very irregular and the boundary is unclear. The shape resembles a map or a lake, which is different from the necrotic morphology of common tumors. Combined with the pathological results, the diffuse patchy low-density lesions may represent the collagen denaturation zone. A cystic lesion in the present study was diagnosed as a true cyst, which is a rare presentation in CDC. Only one of 17 cases was a cystic CDC in a study by Perry et al. . When the essential components and cystic components coexist, careful analysis of the characteristics of the essential component is important for the differential diagnosis.
The majority of CDC tumors in the present cohort were hypovascular. Most of the CDC tumors in the dynamic enhanced scan showed a relatively low density in the renal cortex and medulla. The parenchyma of the mass was uneven, showing light to moderate enhancement in the cortical or medullary phase that was lower than that of the surrounding renal parenchyma. The medullary phase showed uneven and mildly delayed enhancement. The degree of enhancement was lower than that of the renal parenchyma, which is consistent with the results of a Chinese study. This enhancement pattern differed from that of blood-rich clear cell renal cell carcinoma, renal medullary carcinoma, renal angiomyolipoma, and renal angioma. Seong et al.. reported that unlike more common conventional RCC, contrast-enhanced CT scans of CDCs usually show weak (69%) and heterogeneous enhancement (85%). This enhancement pattern differs from that of renal clear cell carcinoma, which shows significant enhancement in the cortical phase, with the density reaching a peak and decreasing markedly during the medullary phase. This enhancement pattern of CDC tumors also contributes to the identification of renal clear cell carcinoma. Fujimoto et al. analyzed the enhancement pattern of RCCs greater than 5 cm in diameter on contrast-enhanced helical CT. They reported that strong enhancement comparable to that of the renal cortex was detected only in conventional RCC (75%). Jeong et al. reported that conventional renal carcinomas show a stronger enhancement than nonconventional renal carcinomas in both the corticomedullary and excretory phases. Tumors with >84 HU in the corticomedullary phase and 44 HU in the excretory phase are likely to be conventional renal carcinomas, whereas the present data showed a 23.5 HU increase in the cortical phase.
In present study, CTA showed that the renal artery was involved in the blood supply and the distal branches were destroyed. The filling defect was observed in the renal vein and inferior vena cava. The tumor thrombus showed an expanded shape and the degree of enhancement was similar to that of the central necrotic and the hypovascular areas. No tumor blood vessels were observed in the vicinity of the tumor, and the original renal blood vessels were not thickened or significantly displaced. These characteristics indicate that CDC tumor cells do not produce angiogenic factors, and the original renal artery branches are rarely destroyed by tumor cells. The fibrous tissue in the tumor stroma can compress the intratumoral vessels; this differs from renal tubular cell carcinoma, which is characterized by vessel hyperplasia and a spherical shape of tumors.
Most RCCs have a low FDG metabolism and are similar to the normal renal parenchyma. 18F-FDG PET has certain limitations in the detection and diagnosis of common renal cancer. Because CDC is rare, there is little information on PET performance in the literature. Ye et al. reported a case of CDC with a maximum diameter of 4.6 cm in the right kidney. The SUVmax of PET was 7.0. Two patients in our group underwent PET/CT examination, and the primary lesions were highly metabolic, with SUVmax values of 14.9 and 14.3, respectively. One of the PET/CT images showed a higher metabolism in the lymph nodes, lungs, pleura, and multiple bone metastases, which was consistent with the HU and other studies. Compared with other common renal cancer pathological types (such as clear cell carcinoma), CDC is characterized by high invasiveness and a poor prognosis, and it frequently shows high FDG uptake. 18F-FDG PET/CT is effective for the diagnosis of renal tumor metastasis. Besne et al. showed that the five-year survival rate of patients with distant metastasis of urinary tumors is 0–20%. However, resection of isolated metastases increases the five-year survival rate to 25–50%. Therefore, early detection of metastases is essential. Safaei et al. reported that the sensitivity and specificity of PET for detecting renal cell carcinoma metastases are 87% and 100%, respectively. Majhail et al. analyzed the biopsy or surgical resection samples from 36 metastatic lesions in 24 patients with RCC. The results showed that the specificity and positive predictive value of 18F-FDG PET/CT for distant metastasis were 100%. In this study, PET/CT of the lymph nodes showed no metastasis after surgical resection. The diagnosis of lymph node metastasis by PET/CT needs to be further investigated.
Local invasion and distant metastasis
CDC is a highly malignant tumor that often shows strong invasiveness and early metastasis[5,7,10,15-16,18]. The incidence of extracranial metastasis of CDC tumors in this group supports the early metastasis of CDC. The incidence of metastasis reached 69%. The rapid metastatic spread and aggressiveness of CDC may be due to its central or perihilar location. CDC is characterized by infiltration into the kidney and local lymph node metastasis, as well as distant metastasis. Most patients show lymph node enlargement and metastasis to distant organs. Lymph node metastasis accounts for 80%, lung and adrenal metastasis account for 25%, and liver metastasis accounts for 20%. The prognosis is extremely poor, and patients die within two years of onset. In this series, nine (69.2%) cases had lymphatic metastasis, three (23.1%) cases had bilateral lung metastasis, two (15.4%) cases had adrenal gland involvement, one (7.7%) case had inferior vena cava involvement, one (7.7%) case had bilateral pleural metastasis, one (7.7%) case had brain metastasis, one (7.7%) case had bone metastases (including the right rib, pubis, and left scapula). This is due to its high degree of malignancy and invasive biological characteristics.
The present study had several limitations. The main limitation was that the number of CDCs was too small for the CT and histopathologic analysis to be significant. Further studies with a larger number of cases are necessary.