Diagnostic value of contrast-enhanced CT in clear cell renal cell carcinoma: a systematic review and meta-analysis

Abstract

Objective

Contrast-enhanced computed tomography (CECT) improves lesion contrast with surrounding tissues through the injection of contrast agents. This enhancement allows for more precise lesion characterization, aiding in the early diagnosis of clear cell renal cell carcinoma (ccRCC). This meta-analysis aims to assess the diagnostic efficacy of CECT in ccRCC and to provide an ideal imaging examination method for the preoperative diagnosis of ccRCC.

Methods

We conducted a comprehensive search across six major online databases: PubMed, Web of Science, Cochrane Library, WANFANG DATA, China National Knowledge Infrastructure, and Chinese BioMedical Literature Database (CBM). The objective was to collate and analyze studies that evaluate the diagnostic utility of CECT in the identification of ccRCC. Meta-disc 1.4 and Stata 16.0 were used to conduct a meta-analysis and evaluate the diagnostic accuracy of CECT for ccRCC.

Results

The meta-analysis included 17 relevant studies investigating the diagnostic value of CECT for ccRCC. The combined sensitivity and specificity of CECT were 0.88 (95% confidence interval: 0.83–0.91) and 0.82 (95%CI: 0.75–0.87), respectively. Positive diagnostic likelihood ratio = 4.87 (95%CI: 3.47–6.84), negative diagnostic likelihood ratio = 0.15 (95%CI: 0.11–0.21), and diagnostic odds ratio = 32.67 (95%CI: 18.21–58.61). In addition, the area under the ROC curve was 0.92 (95%CI: 0.89–0.94), indicating that CECT has a decent discriminative ability in diagnosing ccRCC.

Conclusions

CECT is recognized as a highly effective imaging tool for diagnosing ccRCC. It provides valuable guidance in the preoperative assessment and planning of surgical strategies for patients with ccRCC.

Introduction

Renal cell carcinoma (RCC), a prevalent malignancy within the urological spectrum, originates from the renal cortex and constitutes approximately 2% of all adult cancers globally [1]. RCC is classified into two primary subtypes: clear cell renal cell carcinoma (ccRCC) and non-clear cell renal cell carcinoma (non-ccRCC), including chromophobe renal cell carcinoma (ChRCC) and papillary renal cell carcinoma (pRCC). The predominant subtype, ccRCC [2], is known for its high malignancy and poor prognosis, leading to a higher mortality rate. ChRCC and pRCC have distinct pathological characteristics and prognostic outcomes. Characterized by extensive vascularization, ccRCC is a highly blood-supplied tumor, predisposing it to aggressive growth and distant metastasis [3]. The average five-year survival rate for ccRCC hovers around 55–60%. Notably, research indicates that patients with localized RCC lesions, absent peripheral invasion or distant metastasis, exhibit a significantly higher five-year survival rate of 91.7% [4]. This underscores the critical importance of early diagnosis in enhancing patient survival prospects.

Percutaneous renal biopsy stands as the definitive standard for the preoperative diagnosis of ccRCC. However, performing a preoperative biopsy requires careful consideration of factors such as the risk of tumor spread, the size of the biopsy needle, the physician’s proficiency, and the potential impact on treatment strategies [5]. Early stages of ccRCC often lack distinct clinical symptoms, leading to most diagnoses occurring incidentally during imaging investigations [6]. Imaging modalities provide a non-invasive, safe means to visualize internal body structures, organs, and densities, playing a crucial role in both the primary diagnosis and differential diagnosis of ccRCC. Consequently, the diagnostic process for ccRCC typically commences with an anatomical evaluation using imaging techniques, followed by histopathological validation.

Computed tomography (CT) stands as a primary imaging technique in the diagnosis of renal tumors, offering widespread availability and precise visualization of tumor extent. CECT, an advanced form of standard CT, is acknowledged as the gold standard in renal tumor imaging. In ccRCC, CT typically reveals lesions with heterogeneous iso- or hypo-densities compared to the normal renal parenchyma. A characteristic feature of ccRCC in CECT is a rapid enhancement followed by a quick washout, with lesions reaching a rapid peak in CT values during the corticomedullary phase and displaying moderate-to-high enhancement, followed by a swift decline in lesion density in the nephrographic phase, often lower than that of the surrounding renal parenchyma. A study involving 170 ccRCC, 57 pRCC, and 22 chRCC cases indicated a peak enhancement of ccRCC during the corticomedullary phase [7]. Liang and colleagues [8] analyzed CECT features in 82 ccRCC, 24 pRCC, and 19 chRCC cases, noting that ccRCC generally shows significantly greater contrast enhancement compared to most pRCC and chRCC. Heterogeneous enhancement is more frequently observed in ccRCC and pRCC lesions, whereas chRCC lesions more commonly exhibit uniform enhancement. Corroborating these findings, a retrospective study in the United States highlighted the significantly higher enhancement levels in ccRCC compared to other renal tumor types, with ccRCC predominantly showing heterogeneous enhancement in the nephrographic phase [9].

CECT can observe the changes in kidney lesions in real time through multiple directions and sections, and can also clearly display the surrounding blood flow conditions, providing a reference for the diagnosis of ccRCC before surgery. This study involved a comprehensive review of the literature on the use of CECT in the diagnosis of ccRCC. A meta-analysis was subsequently performed to demonstrate the high diagnostic efficacy of CECT in identifying ccRCC, providing a reliable imaging assessment for preoperative planning.

Methods

This study has been registered on PROSPERO with the registration number [CRD42024555363].

Search strategy

We conducted a comprehensive search across six online databases: PubMed, Web of Science, Cochrane Library, WANFANG DATA, China National Knowledge Infrastructure (CNKI), and the Chinese BioMedical Literature Database (CBM). The search covered all publications up to April 31, 2024. Our search strategy employed a combination of controlled vocabulary and free-text terms, tailored to each database’s unique features. The primary keywords included “contrast-enhanced computed tomography,” “CECT,” “clear cell renal cell carcinoma,” “ccRCC,” “computed tomography,” “enhanced CT,” “renal cell carcinoma, clear cell,” and “kidney cancer, clear cell.” To ensure comprehensive coverage, we also examined the references of the retrieved articles and reviewed relevant meta-analyses and papers. Additionally, a manual search was conducted to address any potential gaps in automated searches.

Inclusion and exclusion criteria

The evaluation process was conducted independently by two researchers, who thoroughly reviewed the titles, abstracts, and full texts of retrieved studies, strictly adhering to the established inclusion and exclusion criteria. In instances of disagreement regarding the inclusion of a study, a determination was made by discussion. The inclusion criteria for this study were defined as follows: (1) Literature related to the diagnosis of ccRCC using CECT, limited to publications in English and Chinese. (2) Diagnostic experimental studies. (3) Studies where the gold standard for diagnosing ccRCC involved pathological outcomes or long-term imaging follow-up. (4) Studies that provided complete datasets. The study established the following exclusion criteria: (1) Studies where the full text was not accessible. (2) Literature not about diagnostic experiments, including conference proceedings, lectures, case reports, and abstracts. (3) Studies that did not report essential diagnostic metrics, specifically true positive (TP), false positive (FP), false negative (FN), and true negative (TN) values. (4) Documents that represented redundant publications of previously released material.

Data extraction and quality assessment

Two researchers (SJC and ZYH) independently assessed the quality of the literature and extracted the data. Both researchers hold advanced degrees in medical research and have extensive experience in conducting systematic reviews and meta-analyses, with multiple peer-reviewed publications in the field of medical imaging and oncology. In cases of disagreement, issues were resolved through discussion. Using a specifically designed data extraction form, two independent researchers undertook the task of extracting data from the selected studies, subsequently creating a comprehensive database. This database included pertinent details from each study, such as the first author’s name, publication year, country where the study was executed, study type, patient demographics (number, gender, age), the specific contrast agent and its dosage used in CECT, diagnostic methodologies, and key diagnostic values including true positives, false positives, false negatives, and true negatives.

The quality of each study included in our analysis was thoroughly evaluated using the 14-item Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) instrument [10]. Each item on this scale was assessed with one of three possible responses: ‘Yes’, ‘No’, or ‘Unclear’.

Statistical analysis

Data extracted from the identified studies were analyzed using Meta-Disc 1.4 and Stata 16.0 software [11, 12]. We assessed the diagnostic efficacy of CECT in detecting ccRCC through metrics such as sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), diagnostic odds ratio (DOR), and the area under the ROC curve (AUC). Initially, Meta-Disc 1.4 was employed for threshold effect testing to determine heterogeneity among the included studies. The Cochrane Q test and I² statistic were then applied to evaluate inter-study heterogeneity, quantifying it based on P-values and I². Further analysis of effect sizes, sensitivity, specificity, PLR, NLR, DOR, AUC, and their 95% confidence intervals (CIs) was conducted using Stata 17.0. A fixed-effect model was adopted for meta-analysis in cases of low heterogeneity (P > 0.05 and I² < 50%), while a random-effects model was used when heterogeneity was significant (P ≤ 0.05 or I² ≥ 50%). Sensitivity analyses were also performed to examine the robustness and reliability of the aggregated findings. Lastly, publication bias was investigated using Deek’s funnel plot, with a P-value ≤ 0.05 suggesting potential bias.

Results

Literature selection process, characteristics and quality assessment of included studies

In our comprehensive search across six databases, we identified a total of 4,167 articles: PubMed (678), Cochrane Library (11), Web of Science (237), China National Knowledge Infrastructure (2,095), WanFang Data (813), and CBM (333). No further articles were procured through manual search. After removing 2,913 duplicates, we reviewed the titles and abstracts, narrowing the selection to 20 articles. After full-text evaluation, three articles were excluded due to data unavailability or unusable format. Consequently, 17 articles were included in the meta-analysis (Fig. 1), involving 1,370 patients. The essential characteristics of the 12 studies included in the meta-analysis post-screening are presented in Table 1. [8, 13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28].

Table 1 Baseline characteristics of included studies for meta-analysis

Fig. 1

PRISMA 2020 flow diagram for new systematic reviews

The 17 articles incorporated into this study underwent evaluation using the QUADAS-2 scale, a diagnostic quality assessment tool with 14 criteria. Each criterion was assessed in relation to its relevance to the included studies, employing “Yes,” “No,” and “Unclear” as response options. Table 2 presents the outcome of this quality assessment. All studies in this analysis were benchmarked against a gold standard. Criteria 1, 2, 3, 5, 6, 7, 8, 9, and 10 were assessed as “Yes,” demonstrating that these studies adhered to reference standards consistent with the gold standard, thus minimizing bias. Criterion 4 was categorized as “Low Risk,” indicating a negligible bias in case selection. Criterion 11 received a consistent “Yes” rating, signifying the appropriateness of the interval between the evaluation test and the gold standard in the studies. Criteria 12, 13, and 14 were rated as “No,” suggesting a significant potential for bias.

Table 2 Details of quality assessment with QUADAS scale diagnostic quality evaluation form

Heterogeneity test

Utilizing Meta-DiSc 1.4 software, we determined a Spearman correlation coefficient of -0.028 (P = 0.914), suggesting no threshold effect in the diagnostic methodology of CECT. Heterogeneity assessments carried out via Stata 16.0 showed significant variation for sensitivity, specificity, PLR, NLR, and DOR, as indicated by Cochran-Q tests yielding P-values less than 0.05. Moreover, the I² values for these indicators were predominantly above 50% (as shown in Supplemental Figs. 1, 2, and 3). Consequently, a random-effects model was utilized in this study to integrate these diverse effect sizes.

Pooling of effect sizes for diagnostic value assessment

This study incorporated 17 articles that evaluated the diagnostic efficacy of CECT in identifying ccRCC. The synthesized results are detailed in Table 3. The aggregated sensitivity was found to be 0.88 (95% CI: 0.83–0.91), and specificity was 0.82 (95% CI: 0.75–0.87). The PLR stood at 4.87 (95% CI: 3.47–6.84), while the NLR was 0.15 (95% CI: 0.11–0.21). The DOR was calculated as 32.67 (95% CI: 18.21–58.61). Additionally, the AUC of the summary receiver operating characteristic (SROC) was 0.92 (95% CI: 0.89–0.94). An AUC value nearing 1 signifies the high diagnostic accuracy of CECT in ccRCC, indicating the method’s robustness (Fig. 2).

Table 3 Pooled effect size for diagnostic value assessment

Fig. 2

The area under the curve of SROC in CECT

Publication bias analysis

The Deeks funnel plot analysis of the 17 studies included in the meta-analysis demonstrated an even distribution of DOR values on both sides of the pooled effect size (p = 0.48), indicating the absence of significant publication bias among the included studies in the meta-analysis (Fig. 3).

Fig. 3

Deeks’ funnel plot for publication assessment

Evaluation of clinical effect

Applying Fagan’s nomogram, we assessed the clinical utility of CECT in the diagnosis of ccRCC. With an initial pre-test probability of 50%, the post-test probability escalated to 83%. Conversely, maintaining the same pre-test probability at 50%, the post-test probability was markedly reduced to 13%. This analysis substantiates the significant clinical relevance of CECT in accurately diagnosing ccRCC (Fig. 4).

Fig. 4

Fagan’s nomogram in detecting diagnostic probability of CECT for ccRCC

Discussion

This study builds on the previous systematic review which included 40 articles analyzing various imaging modalities such as computed tomography, magnetic resonance imaging (MRI), positron emission tomography-CT (PET-CT), and ultrasound (US) for diagnosing and staging renal-cell carcinoma in adults. While the previous review provided a broad assessment of multiple imaging techniques for RCC in a cohort of 4354 patients, our study specifically narrows the focus to the diagnostic performance of CECT in ccRCC. By concentrating solely on CECT, we aim to provide a more detailed and precise analysis of its diagnostic accuracy in identifying and staging ccRCC. This focused approach allows us to demonstrate CECT’s robustness and effectiveness as a diagnostic tool for ccRCC, providing specific insights into its utility that were not the primary focus of the broader meta-analysis. In this analysis, 17 studies with a collective cohort of 1,370 patients were reviewed. The sensitivity and specificity of CECT for diagnosing ccRCC exceeded 80%, underscoring CECT’s robust diagnostic performance. Additionally, a higher PLR, a lower NLR, and a DOR exceeding 1, especially with increasing values, signify enhanced diagnostic and differential diagnostic proficiency. In this context, CECT’s PLR of 4.87, NLR of 0.15, and DOR of 32.67 for ccRCC diagnosis reinforce its substantial diagnostic and differential diagnostic strength. Moreover, the AUC for CECT stood at 0.92, nearing 1, further indicating its high accuracy and efficacy in the diagnosis of ccRCC.

Several studies have highlighted the distinct advantages of CECT in diagnosing ccRCC [29,30,31]. CECT is unaffected by factors such as breathing, body size, or acoustic shadows from gas and bones, enabling clear anatomical cross-sections. Its extensive scanning range allows for the effective evaluation of potential distant metastases. Additionally, CECT can assess enhancement in multiple lesions within a single kidney and detect abnormal lesions in the opposite kidney, offering a broader diagnostic scope.

In addition, research has established that the pathological grade of ccRCC serves as an independent prognostic factor [32]. The Fuhrman grading system is instrumental in determining the malignancy severity, metastatic potential, and aggressive behavior of renal cancer. A higher Fuhrman grade in ccRCC correlates with increased malignancy and aggressiveness, leading to decreased survival rates. In contrast, lower-grade ccRCC, characterized by lower malignancy levels, can be managed with diverse treatment options, generally resulting in a positive prognosis [32]. Yang et al. demonstrated that higher Fuhrman grades, as identified through triphasic dynamic contrast-enhanced CT scans in ccRCC, are associated with lower cancer-specific and 3-year survival rates [33]. Furthermore, pathological grading, a key diagnostic tool for ccRCC, manifests distinct enhancement levels in dynamic contrast-enhanced CT scans [34]. Studies have consistently shown that lower-grade ccRCC (Fuhrman grades 1–2) display higher enhancement values in preoperative CT scans compared to their higher-grade counterparts (Fuhrman grades 3–4) [35, 36]. Additionally, CT imaging features such as necrosis can be an independent predictor of higher-grade ccRCC, as per the Fuhrman grading system [37]. Coy et al. [35], in their study of 127 cases of ccRCC, found that the enhancement values of 3D volumetric CT during the corticomedullary and excretory phases negatively correlated with tumor grade. Notably, the enhancement was significantly higher in lower-grade lesions than in higher-grade ones. This finding was potentially linked to the increased necrosis and the diminished, less mature vasculature characteristic of higher-grade lesions.

This study is subject to several limitations. First, a high degree of heterogeneity was observed in the majority of effect sizes analyzed. Subgroup analyses were not conducted, and factors such as the type of contrast agent and lesion size may have contributed to this heterogeneity. Second, the study populations in the majority of included articles were predominantly Asian, raising the possibility of regional bias. Third, of the primary studies included in our analysis, only one was a prospective study. The predominance of retrospective studies may lead to an overestimation of the accuracy of diagnostic tests. Consequently, there is a need for further prospective research to more accurately assess the diagnostic efficacy of CECT in ccRCC.

Conclusion

Our study demonstrates that CECT has high sensitivity and specificity for diagnosing ccRCC. Future research comparing CECT with other imaging modalities like MRI, PET-CT, and US is needed to establish the most effective imaging strategy for ccRCC. CECT’s significant utility in the preoperative assessment of ccRCC makes it valuable for evaluating patient conditions and planning surgical strategies.