Infiltrative Renal Malignancies: Imaging Features, Prognostic Implications, and Mimics

Differentiating between renal medullary and clear cell renal carcinoma with a machine learning radiomics approach

BACKGROUND

The objective of this study was to develop and validate a radiomics-based machine learning (ML) model to differentiate between renal medullary carcinoma (RMC) and clear cell renal carcinoma (ccRCC).

METHODS

This retrospective Institutional Review Board -approved study analyzed CT images and clinical data from patients with RMC (n = 87) and ccRCC (n = 93). Patients without contrast-enhanced CT scans obtained before nephrectomy were excluded. A standard volumetric software package (MIM 7.1.4, MIM Software Inc.) was used for contouring, after which 949 radiomics features were extracted with PyRadiomics 3.1.0. Radiomics analysis was then performed with RadAR for differential radiomics analysis. ML was then performed with extreme gradient boosting (XGBoost 2.0.3) to differentiate between RMC and ccRCC. Three separate ML models were created to differentiate between ccRCC and RMC. These models were based on clinical demographics, radiomics, and radiomics incorporating hemoglobin electrophoresis for sickle cell trait, respectively.

RESULTS

Performance metrics for the 3 developed ML models were as follows: demographic factors only (AUC = 0.777), calibrated radiomics (AUC = 0.915), and calibrated radiomics with sickle cell trait incorporated (AUC = 1.0). The top 4 ranked features from differential radiomic analysis, ranked by their importance, were run entropy (preprocessing filter = original, AUC = 0.67), dependence entropy (preprocessing filter = wavelet, AUC = 0.67), zone entropy (preprocessing filter = original, AUC = 0.67), and dependence entropy (preprocessing filter = original, AUC = 0.66).

CONCLUSION

A radiomics-based machine learning model effectively differentiates between ccRCC and RMC. This tool can facilitate the radiologist's ability to suspicion and decrease the misdiagnosis rate of RMC.

Predominantly multiple focal non-cystic renal lesions: an imaging approach

Multiple non-cystic renal lesions are occasionally discovered during imaging for various reasons and poses a diagnostic challenge to the practicing radiologist. These lesions may appear as a primary or dominant imaging finding or may be an additional abnormality in the setting of multiorgan involvement. Awareness of the imaging appearance of the various entities presenting as renal lesions integrated with associated extrarenal imaging findings along with clinical information is crucial for a proper diagnostic approach and patient work-up. This review summarizes the most relevant causes of infectious, inflammatory, vascular, and neoplastic disorders presenting as predominantly multiple focal non-cystic lesions.

Left ovarian mass revealing multivisceral lymphoma

Lymphoma encompasses a range of cancers originating in the lymphatic system, categorized into Hodgkin lymphoma and non-Hodgkin lymphoma. Hodgkin lymphoma classically present as nodal disease, whereas non-Hodgkin lymphoma tends to involve extranodal regions. While it can be part of a systemic lymphoma, isolated nodal involvement is not uncommon. Extranodal lymphoma can affect virtually any organ or tissue, with the spleen, liver, gastrointestinal tract, pancreas, abdominal wall, genitourinary tract, adrenal glands, peritoneal cavity, and biliary tract being among the most commonly involved sites, in decreasing order of frequency. We present a case involving a 54-year-old woman presented with left iliac fossa pain. A sonography was performed, which showed left pelvic mass, magnetic resonance imaging showed left ovarian mass with enlargement of the cervix. Computed tomography revealed enlargement of the pancreas and adrenal glands, along with masses in the kidneys associated with extensive pathological lymph node enlargement in the para-aortic and pelvic regions. The patient underwent biopsy of a para-aortic lymph node, which revealed a diffuse large B cell lymphoma.

Prognostic Impact and Genomic Backgrounds of Renal Parenchymal Infiltration or Micronodular Spread in Nonmetastatic Clear Cell Renal Cell Carcinoma

A subset of clear cell renal cell carcinomas (ccRCCs) exhibits various growth patterns that infiltrate the normal renal parenchyma; however, our understanding of its association with cancer aggressiveness is incomplete. Here, we show that the morphology of the tumor interface with normal renal parenchyma is robustly associated with cancer recurrence after surgery, even when compared with the TNM staging system or the World Health Organization/International Society of Urological Pathology (WHO/ISUP) nuclear grade in nonmetastatic ccRCC. Hematoxylin and eosin-stained slides of whole tissue sections from surgical specimens were analyzed using a cohort of 331 patients with nonmetastatic ccRCC treated with radical nephrectomy. The patients were classified into 10 subgroups based on our classification algorithms for assessing the tumor interface with normal renal parenchyma. Among the 10 subgroups, 4 subgroups consisting of 40 patients (12%) were identified to have aggressive forms of nonmetastatic ccRCC associated with poor prognosis and unified as renal parenchymal infiltration or micronodular spread (RPI/MNS) phenotypes. Multivariable analyses showed that RPI/MNS phenotypes were robustly associated with shorter disease-free survival, independently of existing pathological factors including the TNM staging system and WHO/ISUP nuclear grade. The hazard ratio was highest for RPI/MNS (4.62), followed by WHO/ISUP grades 3 to 4 (2.11) and ≥pT3a stage (2.05). In addition, we conducted genomic analyses using next-generation sequencing of infiltrative lesions in 18 patients with RPI/MNS and tumor lesions in 33 patients without RPI/MNS. Results showed that alterations in SETD2 and TSC1 might be associated with RPI/MNS phenotypes, whereas alterations in PBRM1 might be associated with non-RPI/MNS phenotypes. These data suggest that RPI/MNS may be associated with aggressive genomic backgrounds of ccRCC, although more comprehensive analyses with a larger sample size are required. Future studies may further elucidate the clinical implications of RPI/MNS, particularly for deciding the indication of adjuvant treatment after nephrectomy.

Imaging features of perinephric myxoid pseudotumors of fat

OBJECTIVE

Retrospectively evaluate multimodality imaging features of perinephric myxoid pseudotumor of fat (PMPTF).

METHODS

Institutional cases of PMPTF with CT, MRI and/or ultrasound evaluation from 1/1/2020 to 9/1/2023 were retrospectively reviewed. Patient demographics and clinical history were reviewed, and imaging features recorded.

RESULTS

14 patients with pathologically-proven PMPTF were identified (11 M, 3 F; mean age 66.7 ± 17.0 years; range 40-87 years). Three patients (18%) had bilateral lesions; a total of 17 PMPTFs were reviewed. 15/17 (88%) were biopsy-proven; two cases were diagnosed by imaging only in patients with a contralateral biopsy-proven PMPTF. All evaluable specimens were negative for MDM2 amplification. 11/17 (65%) occurred in patients with renal disease, including 4/17 (24%) in patients with renal transplant. 100% (17/17) had CT, 11/17 (65%) MRI, and 6/17 (35%) ultrasound. The mean largest lesion dimension was 10.9 ± 4.6 cm (range 4.3-17.0 cm). Of cases involving native kidneys, 7/13 (54%) presented as multifocal perinephric masses and 5/13 (38%) as a solitary perinephric mass. All four transplant cases presented as infiltrative-appearing masses involving the renal sinus with lesser perinephric involvement. 14/17 (82%) lesions contained macroscopic fat on CT and MRI and 3/17 (18%) showed no macroscopic fat, all involving renal transplants. All cases with MRI demonstrated T2 hyperintensity with signal dropout on opposed-phase imaging. 11/13 (85%) PMPTF showed no or equivocal CT enhancement. Enhancement was better seen on MRI in all cases evaluated by both CT and MRI. Of the six PMPTFs imaged by ultrasound, four (67%) were heterogeneously hypoechoic and two (33%) had mixed regions of hypo-, iso- and hyperechogenicity relative to adjacent renal parenchyma.

CONCLUSIONS

PMPTF is a rare, benign, and underrecognized lesion that may mimic malignancy, particularly retroperitoneal well-differentiated liposarcoma. The imaging features of this unusual pseudosarcoma differ in native and transplanted kidneys. Improved awareness of this entity will facilitate appropriate patient management and avoid unnecessary intervention.

Deep learning system for malignancy risk prediction in cystic renal lesions: a multicenter study

OBJECTIVES

To develop an interactive, non-invasive artificial intelligence (AI) system for malignancy risk prediction in cystic renal lesions (CRLs).

METHODS

In this retrospective, multicenter diagnostic study, we evaluated 715 patients. An interactive geodesic-based 3D segmentation model was created for CRLs segmentation. A CRLs classification model was developed using spatial encoder temporal decoder (SETD) architecture. The classification model combines a 3D-ResNet50 network for extracting spatial features and a gated recurrent unit (GRU) network for decoding temporal features from multi-phase CT images. We assessed the segmentation model using sensitivity (SEN), specificity (SPE), intersection over union (IOU), and dice similarity (Dice) metrics. The classification model's performance was evaluated using the area under the receiver operator characteristic curve (AUC), accuracy score (ACC), and decision curve analysis (DCA).

RESULTS

From 2012 to 2023, we included 477 CRLs (median age, 57 [IQR: 48-65]; 173 men) in the training cohort, 226 CRLs (median age, 60 [IQR: 52-69]; 77 men) in the validation cohort, and 239 CRLs (median age, 59 [IQR: 53-69]; 95 men) in the testing cohort (external validation cohort 1, cohort 2, and cohort 3). The segmentation model and SETD classifier exhibited excellent performance in both validation (AUC = 0.973, ACC = 0.916, Dice = 0.847, IOU = 0.743, SEN = 0.840, SPE = 1.000) and testing datasets (AUC = 0.998, ACC = 0.988, Dice = 0.861, IOU = 0.762, SEN = 0.876, SPE = 1.000).

CONCLUSION

The AI system demonstrated excellent benign-malignant discriminatory ability across both validation and testing datasets and illustrated improved clinical decision-making utility.

CRITICAL RELEVANCE STATEMENT

In this era when incidental CRLs are prevalent, this interactive, non-invasive AI system will facilitate accurate diagnosis of CRLs, reducing excessive follow-up and overtreatment.

KEY POINTS

The rising prevalence of CRLs necessitates better malignancy prediction strategies. The AI system demonstrated excellent diagnostic performance in identifying malignant CRL. The AI system illustrated improved clinical decision-making utility.

CT-based radiomics signature for differentiating pyelocaliceal upper urinary tract urothelial carcinoma from infiltrative renal cell carcinoma

Objectives

To develop a CT-based radiomics model and a combined model for preoperatively discriminating infiltrative renal cell carcinoma (RCC) and pyelocaliceal upper urinary tract urothelial carcinoma (UTUC), which invades the renal parenchyma.

Materials and methods

Eighty patients (37 pathologically proven infiltrative RCCs and 43 pathologically proven pyelocaliceal UTUCs) were retrospectively enrolled and randomly divided into a training set (n = 56) and a testing set (n = 24) at a ratio of 7:3. Traditional CT imaging characteristics in the portal venous phase were collected by two radiologists (SPH and ZXL, who have 4 and 30 years of experience in abdominal radiology, respectively). Patient demographics and traditional CT imaging characteristics were used to construct the clinical model. The radiomics score was calculated based on the radiomics features extracted from the portal venous CT images and the random forest (RF) algorithm to construct the radiomics model. The combined model was constructed using the radiomics score and significant clinical factors according to the multivariate logistic regression. The diagnostic efficacy of the models was evaluated using receiver operating characteristic (ROC) curve analysis and the area under the curve (AUC).

Results

The RF score based on the eight validated features extracted from the portal venous CT images was used to build the radiomics model. Painless hematuria as an independent risk factor was used to build the clinical model. The combined model was constructed using the RF score and the selected clinical factor. Both the radiomics model and combined model showed higher efficacy in differentiating infiltrative RCC and pyelocaliceal UTUC in the training and testing cohorts with AUC values of 0.95 and 0.90, respectively, for the radiomics model and 0.99 and 0.90, respectively, for the combined model. The decision curves of the combined model as well as the radiomics model indicated an overall net benefit over the clinical model. Both the radiomics model and the combined model achieved a notable reduction in false-positive and false-negativerates, resulting in significantly higher accuracy compared to the visual assessments in both the training and testing cohorts.

Conclusion

The radiomics model and combined model had the potential to accurately differentiate infiltrative RCC and pyelocaliceal UTUC, which invades the renal parenchyma, and provide a new potentially non-invasive method to guide surgery strategies.

Typical and Atypical Imaging Features of Malignant Lymphoma in the Abdomen and Mimicking Diseases

Malignant lymphoma typically presents with homogeneous enhancement of enlarged lymph nodes without internal necrotic or cystic changes on multiphasic CT, which can be suspected without invasive diagnostic methods. However, some subtypes of malignant lymphoma show atypical imaging features, which makes diagnosis challenging for radiologists. Moreover, there are several lymphoma-mimicking diseases in current clinical practice, including leukemia, viral infections in immunocompromised patients, and primary or metastatic cancer. The ability of diagnostic processes to distinguish malignant lymphoma from mimicking diseases is necessary to establish effective management strategies for initial radiological examinations. Therefore, this study aimed to discuss the typical and atypical imaging features of malignant lymphoma as well as mimicking diseases and discuss important diagnostic clues that can help narrow down the differential diagnosis.

Prognostic impact of the radiological infiltrative feature of primary renal tumor in metastatic renal cell carcinoma

OBJECTIVES

Recent studies suggest that the radiological infiltrative feature (r-IF) of renal tumors is strongly correlated with poor oncologic outcomes in locally advanced renal cell carcinoma (RCC). This study investigated the prognostic impact of r-IF of primary renal tumors in metastatic RCC (mRCC) in comparison with International Metastatic RCC Database Consortium (IMDC) risk model.

METHODS

We retrospectively analyzed 91 patients with previously untreated mRCC. Dynamic computed tomography of the primary renal tumor was reviewed to assess r-IF, defined as a focally/extensively ill-defined tumor interface with normal renal parenchyma.

RESULTS

The median age was 67 years, and 69 patients (76%) were men. Prior nephrectomy was performed in 47 patients (52%). The median size of the primary renal tumor was 6.7 cm, and 50 patients (55%) presented with cT3-4 stage. Overall, 25 (28%)/52 (57%)/14 (15%) patients were classified into IMDC favorable/intermediate/poor-risk groups, respectively. An image review identified r-IFs in the primary renal tumor in 40 patients (44%). The incidences of r-IFs were 28%/46%/64% in IMDC favorable/intermediate/poor-risk groups, respectively. During a median follow-up of 2.6 years, 31 patients (34%) died of RCC. On multivariable analysis, r-IF and IMDC intermediate-poor risks were independently associated with poor cancer-specific survival (CSS). Two-year CSS were 64%/87% in patients with/without r-IF, respectively. C-index was improved from 0.73 to 0.81 by adding r-IF to the IMDC risk factors.

CONCLUSIONS

R-IF of the primary renal tumor was an independent risk factor for poor CSS in patients with mRCC, which may improve the prognostic accuracy when combined with the IMDC risk model.

Immunotherapy-related renal toxicity causes reversible renal enlargement

PURPOSE

Prior case reports have noted an increase in renal size and perinephric stranding accompanying immunotherapy-related renal toxicity due to checkpoint-inhibitor therapy. The purpose of this investigation was to systematically evaluate if immunotherapy-related renal toxicity affects renal size and possible associated imaging findings.

METHODS

This retrospective multi-hospital study included 25 patients (13 men), mean age 67 years (range 46-83) who received immune-checkpoint inhibitors for cancer treatment, developed biopsy-proven immunotherapy-related nephritis, and who also had abdominal imaging before, during, and after nephritis was diagnosed. Long axis renal diameter, renal corticomedullary differentiation/enhancement and perinephric stranding were evaluated by two readers at three timepoints: (1) prior to checkpoint inhibitor therapy (baseline), (2) after biopsy-proven immunotherapy-related nephritis (post-treatment), and (3) following renal function recovery (follow-up). Intraclass correlation coefficient and Cohen's Kappa were calculated to quantify agreement. Logistic regression analysis was implemented to measure the association between each timepoint and imaging features.

RESULTS

Reader agreement on kidney measurements was excellent (ICC = 0.87). There was an increase in renal size between baseline and post-treatment (p = 0.001), followed by a decrease between post-treatment to follow-up (p < 0.001). Agreement was perfect for abnormal renal corticomedullary differentiation/enhancement (Kappa = 1, p < 0.001) and almost perfect for perinephric stranding (Kappa = 0.97, p < 0.001). Neither post-treatment nor follow-up imaging findings were significantly associated with these findings compared to the baseline (p = 0.2-0.6).

CONCLUSION

Immunotherapy-related renal toxicity was associated with an increase in renal size coincident with acute renal dysfunction.

Imaging Findings in Immunotherapy-related Renal Toxicity

Immunotherapy-related adverse events (irAEs) associated with immune-checkpoint inhibitors can affect nearly any organ system including commonly the luminal gastrointestinal tract, hepatobiliary system, lungs, endocrine glands, and skin, many of which have described imaging manifestations. In patients without clinically suspected irAEs, imaging findings may be the first indication of an abnormality that prompts further workup to facilitate early detection and initiation of appropriate treatment, such as therapy discontinuation or corticosteroid therapy. While some irAEs have well described imaging correlates, such as pneumonitis, hypophysitis, and colitis, others are not well described, such as nephritis. We report 2 cases of irAE nephritis associated with PD-1 inhibitor therapy and their imaging features.

Infiltrative tumor interface with normal renal parenchyma in locally advanced renal cell carcinoma: Clinical relevance and pathological implications

OBJECTIVES

Locally advanced renal cell carcinoma is considered clinically aggressive, despite heterogeneity in survival outcomes. We investigated the clinical relevance and pathological implications of infiltrative tumor interface with normal renal parenchyma on preoperative imaging in locally advanced renal cell carcinoma.

METHODS

A total of 77 patients with locally advanced renal cell carcinoma (≥pT3a Nany M0) who underwent radical or partial nephrectomy (2008-2018) were analyzed. Preoperative dynamic computed tomography images were reviewed to assess radiological infiltrative features. A radiological infiltrative feature was defined as an ill-defined tumor interface with normal renal parenchyma. The tumor interfaces were analyzed histologically and compared with radiological findings.

RESULTS

The median tumor size was 6.4 cm. Lymphadenopathy was observed in four patients (5.2%). Clear cell renal cell carcinoma was diagnosed in 66 patients (86%) and Fuhrman grade was 3-4 in 38 patients (49%). A total of 30 patients (39%) showed radiological infiltrative features, which were significantly associated with larger tumor size and higher clinical T stage. The specificity and sensitivity of radiological infiltrative features in predicting pathological renal parenchymal infiltration were 90 and 64%, respectively. During a median follow-up period of 3.8 years, 27 patients (35%) developed cancer recurrences, and six patients (7.8%) died of renal cell carcinoma. Multivariable analysis showed that the presence of radiological infiltrative features was an independent risk factor for cancer recurrence. Cancer recurrence and cancer-specific mortality were significantly stratified by the presence or absence of radiological infiltrative features (P < 0.001 and P = 0.02, respectively).

CONCLUSIONS

Locally advanced renal cell carcinoma can show radiological infiltrative features preoperatively, which are significantly associated with unfavorable prognosis. Radiological infiltrative features on preoperative imaging correspond with a high specificity to pathological renal parenchymal infiltration.