In this study, we found that unilocular cRCCs has a lower pathological stage and Fuhrman grade and a better prognosis than purely solid RCCs.
The cystic component in RCC has been the focus of scholars for a long time. Hartman first reported unilocular cRCCs and divided cRCCs into 4 categories: 1) intrinsic multiloculated growth; 2) intrinsic unilocular growth; 3) cystic necrosis; 4) origin from the epithelial lining of a preexisting simple cyst [1].
However, this classification has not been widely accepted by pathologists. Cystic necrosis, for example, is actually caused by the imbalance of tumor growth and insufficient blood supply. It is not only found in various pathological types of RCC, but also common in other malignant tumors. The existence of malignant transformation of simple cysts is still debatable and there are rare related reports. Therefore, many pathologists believe that the above two subtypes should be excluded from the subcategories of cRCC, which means that cRCC should only include the unilocular cystic and the multilocular cystic [19,20].
With the continuous increase of related researches, multilocular cRCC has gradually become the focus. Both the 2004 revised and 2016 WHO guidelines only defined multilocular cRCC without mentioning unilocular cRCC [21,4]. Although unilocular cRCC has been reflected in some studies, there is no specific study with a large sample size. Meanwhile, many studies did not emphasize the classification of cRCC subtypes. As a result, the two concepts of continuously redefined “MCRNLMP” and “cystic renal cell carcinoma” have been implemented without coming into conflict, which has caused a lot of confusion for researchers and clinicians.
In some studies, it was found that there were many papillary RCCs in unilocular cRCCs [1,22]. This is consistent with our findings. Group A had significantly more papillary RCC than group B, which is somewhat inconsistent with the findings of previous studies in that, the main histology of unilocular cRCCs in our study is clear cell carcinoma, not papillary RCCs. This discrepancy may be due to the small number of samples used in previous studies. At the same time, Hartman also noticed that a few cysts discovered on CT imaging are actually extensive necrosis and tumor necrosis is one of the negative indicators in many prognostic models of RCC [23]. In this study, we defined unilocular cRCC, which excludes necrotic cysts through surgical pathology. However, it is worth noting that it is not reliable to determine whether the cystic component is a true cyst or a necrotic cyst only though preoperative CT imaging. For instance, radiologists with more than 3 years of work experience reviewed the imaging data in this study, but among the 83 cases initially judged as true cysts, 9 cases (11%) were still pathologically confirmed as necrotic cysts. Some studies have shown that MRI may be more advantageous than CT in the differential diagnosis of true cysts [24] but due to lack of relevant imaging data, we did not investigate that.
Numerous studies on MCRNLMP have shown that the malignancy of this type of tumor is low, and no recurrence or metastasis have been reported so far [2,3]. Therefore, its definition is changed from “cancer” to “neoplasm”. However, of the 74 cases of unilocular cRCCs in our study, 1 patient died of metastasis of the RCC. Although no distant metastasis was found at the time of diagnosis, 3 cases of lymphatic metastasis were found in surgical pathology of lymphadenectomy. It can be seen that although the unilocular cRCC is less malignant than the purely solid RCC, its prognosis is not as excellent as that of the MCRNLMP defined by the WHO in 2016.
In this study, we did not set the cut off for unilocular cRCCs like other studies but included all RCCs with true cystic components. But we found that the cystic proportion of 1 case who dead of RCC in group A was not low (52%), and the cystic proportion of 1 case who had lymphatic metastasis at diagnosis in group A also reached 78%. We believe that unilocular cRCCs even with a high cystic proportion also probably had lymphatic metastasis at the time of diagnosis and recurrence after surgery. Although the presence of cystic components in RCCs might predict lower malignant potential, the correlation between high cystic proportion and the mostly indolent outcome has not been fully proven, studies with larger sample sizes are needed in the future.
In previous studies conducted before and after 2016, MCRNLMP was not clearly distinguished from other types of cystic RCCs [11-15]. Given that the MCRNLMP is a rare type and was differently defined in previous studies, the true incidence of MCRNLMP was uncertain. According to some recent studies, it can be inferred that it accounts for roughly 1% of RCCs [9,10], while cRCCs account for 15% [1,25 ,26]. Defining the MCRNLMP as being similar with all other types of cRCCs and making survival comparison with solid RCC certainly will mask the relatively poor prognosis and relatively high malignancy of other types of cRCCs such as the unilocular cRCCs. As a result, when making clinical decisions, surgeons may underestimate the malignancy of cRCCs other than MCRNLMP. It is a bot puzzling that although the pathological stage and Fuhrman grade of 74 unilocular cRCC cases are relatively low, 3 cases (4.1%) still have lymphatic metastasis at diagnosis and the difference with that of solid RCC (8.8%) is not significant (p=0.200). We speculated that this may be due to the relatively small sample size. It may also be because that 3 cases of tumors were large in size. The average maximum diameter of the tumors in group A was 4.45±1.10cm, while that of 3 cases with lymphatic metastasis were 6.57cm, 6.37cm, and 6.63cm, respectively.
The excellent prognosis of MCRNLMP may suggest that the existence of the cystic component of renal carcinoma is somehow related to a favorable prognosis. According to existing researches, it can be seen that the prognosis of MCRNLMP, unilocular cRCC, and solid RCC shows a decreasing trend [7,16,17]. Therefore, the clinical treatment and follow-up decisions of them should be made pertinently. Many studies have pointed out that for MCRNLMP, nephron sparing method should be used as much as possible [27,28]. In view of the findings of this study, we recommend that unilocular cRCCs should also be treated with nephron sparing approaches when feasible. Radical nephrectomy is still needed when the tumor is large, and regional lymph nodes should be removed if the preoperative imaging shows the sign of enlarged lymph nodes. However, our proposal and specific implementation plan need to be confirmed by further prospective studies.
For most cases, the diagnosis of different subtypes of cRCC based on imaging data ultimately needs to be confirmed by postoperative pathology, which means that clinical decisions on nephron sparing or radical approaches have always been made. Clarifying that the tumor belongs to a specific subtype of cRCC can only be used as a guide for postoperative follow-up or retrieval reoperation. Therefore, it is particular important to judge different subtypes by preoperative imaging, especially to identify MCRNLMP and necrotic cysts. Several relevant studies published may help on this issue, but there is still room for improvement in the accuracy of diagnosis [29].
There are several limitations to our study. 1). This study is retrospective in nature; the pathological features were summarized based on the surgical pathology reports. We were unable to personally obtain tumor specimens for further analysis, such as cyst fluid properties, more comprehensive immunohistochemistry, etc. 2). As a result of the intention to investigate unilocular cRCCs and because the total number and the number of cancer-specific deaths in group A were small, the optimistic cut off related to prognosis could not be calculated. 3). Twenty patients were excluded due to incomplete imaging data, which may have caused selection bias to the outcomes. 4). In order to eliminate confounders, we performed 1:2 PSM in 761 solid RCC cases, and eventually enrolled 148 cases in the control group. Therefore, when performing Cox regression, the Goodness of Fit may be not ideal enough. Although the PSM method adopted to select matching objects is random, the solid RCC group still has the possibility of underrepresentation. However, the characteristics of solid RCC in other similar studies with larger sample size can be corroborated [7,11,12].