Classification Of Renal Tumors Research Articles

Clear Cell Papillary Renal Cell Carcinoma: Case Series with Review of the Literature

Abstract Clear cell papillary renal cell carcinoma (CCPRCC) was included as a separate entity in the World Health Organisation classification of renal tumours in 2016. Immuno-histopathological and genetic characteristics are much known, but still, clinical features and long-term follow-up require more consolidated data. We report three cases of CCPRCC, detected in different clinical settings. The first case was co-incidentally diagnosed on routine imaging, in a follow-up case of open cystolithotripsy. The second case presented with haematuria under evaluation; and the third case presented with mild thrombocytopenia without anaemia or a positive Coombs test. All the cases revealed characteristic histopathological examination and immunohistochemistry. None of the patients showed tumour recurrence/metastases on follow-up. The study holds importance as it presents CCPRCC cases in different clinical scenarios, describes a rarely reported case of thrombocytopenia without associated anaemia/positive Coombs test, and emphasises the need to consider CCPRCC in relevant clinico-radiological settings and immuno-histopathological features, due to its indolent nature and consequently better prognosis.

A Novel Classification System of Renal Hilar Tumors for Surgical Guidance: Technique, Outcome, and Safety.

This study was designed to develop an innovative classification and guidance system for renal hilar tumors and to assess the safety and effectiveness of robot-assisted partial nephrectomy (RAPN) for managing such tumors. A total of 179 patients undergoing RAPN for renal hilar tumors were retrospectively reviewed. A novel classification system with surgical techniques was introduced and the perioperative features, tumor characteristics, and the efficacy and safety of RAPN were compared within subgroups. We classified the tumors according to our novel system as follows: 131 Type I, 35 Type II, and 13 Type III. However, Type III had higher median R.E.N.A.L., PADUA, and ROADS scores compared with the others (all p < 0.001), indicating increased operative complexity and higher estimated blood loss [180.00 (115.00-215.00) ml]. Operative outcomes revealed significant disparities between Type III and the others, with longer operative times [165.00 (145.00-200.50) min], warm ischemia times [24.00 (21.50-30.50) min], tumor resection times [13.00 (12.00-15.50) min], and incision closure times [22.00 (20.00-23.50) min] (all p < 0.005). Postoperative outcomes also showed significant differences, with longer durations of drain removal (77.08 ± 18.16h) and hospitalization for Type III [5.00 (5.00-6.00) d] (all p < 0.05). Additionally, Type I had a larger tumor diameter than the others (p = 0.009) and pT stage differed significantly between the subtypes (p = 0.020). The novel renal hilar tumor classification system is capable of differentiating the surgical difficulty of RAPN and further offers personalized surgical steps tailored to each specific classification. It provides a meaningful tool for clinical practice.

Fumarate hydratase-deficient renal cell carcinoma: an oncology care institutional experience.

Renal cell carcinoma (RCC) accounts for 2% of all cancer cases worldwide, and majority are sporadic. The latest World Health Organization (WHO) classification of renal cell tumors (fifth edition, 2022) has molecularly defined renal tumor entities, which includes fumarate hydratase (FH)-deficient RCC. FH-deficient RCC is an aggressive carcinoma caused by pathogenic alterations in FH gene, seen in 15% of patients with hereditary leiomyomatosis and renal cell cancer syndrome (HLRCC) syndrome. These tumors occur more frequently at a younger age and present at an advanced stage, carrying a dismal prognosis. We report a series of 10 cases of FH-deficient RCC. The mean age was 49.8 years, and all cases presented in advanced stages (III and IV). Morphologically, the cases had varied architectural patterns with characteristic eosinophilic macronucleoli and perinucleolar halo. On immunohistochemistry (IHC), all showed diffuse nucleo-cytoplasmic expression of S-(2-succino)-cysteine (2-SC), with loss of FH in seven cases. FH-deficient RCCs are aggressive neoplasms and can be diagnosed using specific IHC markers (FH and 2-SC). These patients should undergo germline testing for FH gene mutation, genetic counseling, and surveillance of family members.

Sarcomatoid and Rhabdoid Renal Cell Carcinoma: Clinical, Pathologic, and Molecular Genetic Features.

Renal cell carcinoma (RCC) with sarcomatoid and rhabdoid morphologies has an aggressive biological behavior and a typically poor prognosis. The current 2022 WHO classification of renal tumors does not include them as distinct histologic entities but rather as transformational changes that may arise in a background of various distinct histologic types of RCC. The sarcomatoid component shows malignant spindle cells that may grow as intersecting fascicles, which is reminiscent of pleomorphic undifferentiated sarcoma. The rhabdoid cells are epithelioid cells with eccentrically located vesicular nuclei with prominent nucleoli and large intracytoplasmic eosinophilic inclusions. Studies have shown that RCCs with sarcomatoid and rhabdoid differentiation have distinctive molecular features. Sarcomatoid RCC harbors shared genomic alterations in carcinomatous and rhabdoid components, but also enrichment of specific genomic alterations in the sarcomatoid element, suggesting molecular pathways for development of sarcomatoid growth from a common clonal ancestor. Rhabdoid differentiation also arises through clonal evolution although less is known of specific genomic alterations in rhabdoid cells. Historically, treatment has lacked efficacy, although recently immunotherapy with PD-1/PD-L1/CTLA-4 inhibitors has produced significant clinical responses. Reporting of sarcomatoid and rhabdoid features in renal cell carcinoma is required by the College of American Pathologists and the International Collaboration on Cancer Reporting. This manuscript reviews the clinical, pathologic, and molecular features of sarcomatoid RCC and rhabdoid RCC with emphasis on the morphologic features of these tumors, significance of diagnostic recognition, the molecular mechanisms of tumorigenesis and differentiation along sarcomatoid and rhabdoid lines, and advances in treatment, particularly immunotherapy.

Kidney Tumor Classification on CT images using Self-supervised Learning

One of the most common diseases affecting society around the world is kidney tumor. The risk of kidney disease increases due to reasons such as consumption of ready-made food and bad habits. Early diagnosis of kidney tumors is essential for effective treatment, reducing side effects, and reducing the number of deaths. With the development of computer-aided diagnostic methods, the need for accurate renal tumor classification is also increasing. Because traditional methods based on manual detection are time-consuming, boring, and costly, high-accuracy tests can be performed faster and at a lower cost with deep learning (DL) methods in kidney tumor detection (KTD). Among the current challenges regarding artificial intelligence-assisted KTD, obtaining more precise programming information and the capacity to group with high accuracy make clinical determination more vital and bring it to an important point for current treatment in KTD prediction. This encourages us to propose a more effective DL model that can effectively assist specialist physicians in the diagnosis of kidney tumors. In this way, the workload of radiologists can be alleviated and errors in clinical diagnoses that may occur due to the complex structure of the kidney can be prevented. A large amount of data is needed during the training of the developed methods. Although various studies have been conducted to reduce the amount of data with feature selection techniques, these techniques provide little improvement in the classification accuracy rate. In this paper, a masked autoencoder (MAE) is proposed for KTD, which can produce effective results on datasets containing some samples and can be directly fine-tuned and pre-trained. Self-supervised learning (SSL) is achieved through self-distillation (SD), which can be reintroduced into the configuration loss calculation using masked patches. The SD loss on the decoder and encoder outputs’ latent representation is calculated operating SSLSD-KTD. The encoder obtains local attention, while the decoder transfers its global attention to calculate losses. The SSLSD-KTD method reached 98.04 % classification accuracy on the KAUH-kidney dataset, including 8400 samples, and 82.14 % on the CT-kidney dataset, containing 840 samples. By adding more external information to the SSLSD-KTD method with transfer learning, accuracy results of 99.82 % and 95.24 % were obtained on the same datasets. Experimental results have shown that the SSLSD-KTD method can effectively extract kidney tumor features with limited data and can be an aid or even an alternative for radiologists in decision-making in the diagnosis of the disease.

Diagnosing liposarcoma on (peri)-renal mass biopsy: A clinicopathological study of 30 cases.

Classification of renal neoplasms on small tissue biopsies is in increasing demand, and maintaining broad differential diagnostic considerations in this setting is necessary. When evaluating a renal or perirenal tumour biopsy with sarcomatoid morphology, together with sarcomatoid renal cell carcinoma and sarcomatoid urothelial carcinoma as top diagnostic considerations, it is vital to additionally consider the possibility of well-differentiated and de-differentiated liposarcoma. This study reports a series of 30 biopsy samples from sites in or around the kidney collected from four institutions in which the correct diagnosis was either well-differentiated or de-differentiated liposarcoma. The majority (26 of 30, 87%) of lesions were accurately diagnosed on biopsy sampling, all of which incorporated testing for MDM2 by immunohistochemistry (IHC), fluorescence in-situ hybridisation (FISH) or a combination of the two as part of the diagnostic work-up. Tumour expression of MDM2 by IHC without confirmatory FISH analysis was sometimes (30%) sufficient to reach a diagnosis, but demonstration of MDM2 amplification by FISH was ascertained in the majority (57%) of biopsy samples. A diagnosis of de-differentiated liposarcoma was not definitively established until resection in four (13%) patients, as no MDM2 testing was performed on the corresponding pre-operative biopsies. When a retroperitoneal tumour is not clinically suspected, histological consideration of a liposarcoma diagnosis may be overlooked. Implementation of ancillary immunohistochemical and cytogenetic testing can ultimately lead to a definitive diagnosis in this potentially misleading anatomical location.

L1 Cell Adhesion Molecule (L1CAM) Expression and Molecular Alterations Distinguish Low-Grade Oncocytic Tumor From Eosinophilic Chromophobe Renal Cell Carcinoma

Renal low-grade oncocytic tumor (LOT) is a recently recognized renal cell neoplasm designated within the “other oncocytic tumors” category in the 2022 World Health Organization classification system. Although the clinicopathologic, immunohistochemical, and molecular features reported for LOT have been largely consistent, the data are relatively limited. The morphologic overlap between LOT and other low-grade oncocytic neoplasms, particularly eosinophilic chromophobe renal cell carcinoma (E-chRCC), remains a controversial area in renal tumor classification. To address this uncertainty, we characterized and compared large cohorts of LOT (n = 67) and E-chRCC (n = 69) and revealed notable differences between the 2 entities. Clinically, LOT predominantly affected women, whereas E-chRCC showed a male predilection. Histologically, although almost all LOTs were dominated by a small-nested pattern, E-chRCC mainly showed solid and tubular architectures. Molecular analysis revealed that 87% of LOT cases harbored mutations in the tuberous sclerosis complex (TSC)-mTOR complex 1 (mTORC1) pathway, most frequently in MTOR and RHEB genes; a subset of LOT cases had chromosomal 7 and 19q gains. In contrast, E-chRCC lacked mTORC1 mutations, and 60% of cases displayed chromosomal losses characteristic of chRCC. We also explored the cell of origin for LOT and identified L1 cell adhesion molecule (L1CAM), a collecting duct and connecting tubule principal cell marker, as a highly sensitive and specific ancillary test for differentiating LOT from E-chRCC. This distinctive L1CAM immunohistochemical labeling suggests the principal cells as the cell of origin for LOT, unlike the intercalated cell origin of E-chRCC and oncocytoma. The ultrastructural analysis of LOT showed normal-appearing mitochondria and intracytoplasmic lumina with microvilli, different from what has been described for chRCC. Our study further supports LOT as a unique entity with a benign clinical course. Based on the likely cell of origin and its clinicopathologic characteristics, we propose that changing the nomenclature of LOT to “Oncocytic Principal Cell Adenoma of the Kidney” may be a better way to define and describe this entity.

Fully Automated Versions of Clinically Validated Nephrometry Scores Demonstrate Superior Predictive Utility versus Human Scores.

To automate the generation of three validated nephrometry scoring systems on preoperative computerised tomography (CT) scans by developing artificial intelligence (AI)-based image processing methods. Subsequently, we aimed to evaluate the ability of these scores to predict meaningful pathological and perioperative outcomes. A total of 300 patients with preoperative CT with early arterial contrast phase were identified from a cohort of 544 consecutive patients undergoing surgical extirpation for suspected renal cancer. A deep neural network approach was used to automatically segment kidneys and tumours, and then geometric algorithms were used to measure the components of the concordance index (C-Index), Preoperative Aspects and Dimensions Used for an Anatomical classification of renal tumours (PADUA), and tumour contact surface area (CSA) nephrometry scores. Human scores were independently calculated by medical personnel blinded to the AI scores. AI and human score agreement was assessed using linear regression and predictive abilities for meaningful outcomes were assessed using logistic regression and receiver operating characteristic curve analyses. The median (interquartile range) age was 60 (51-68) years, and 40% were female. The median tumour size was 4.2 cm and 91.3% had malignant tumours. In all, 27% of the tumours were high stage, 37% high grade, and 63% of the patients underwent partial nephrectomy. There was significant agreement between human and AI scores on linear regression analyses (R ranged from 0.574 to 0.828, all P < 0.001). The AI-generated scores were equivalent or superior to human-generated scores for all examined outcomes including high-grade histology, high-stage tumour, indolent tumour, pathological tumour necrosis, and radical nephrectomy (vs partial nephrectomy) surgical approach. Fully automated AI-generated C-Index, PADUA, and tumour CSA nephrometry scores are similar to human-generated scores and predict a wide variety of meaningful outcomes. Once validated, our results suggest that AI-generated nephrometry scores could be delivered automatically from a preoperative CT scan to a clinician and patient at the point of care to aid in decision making.

Reclassifying Papillary, Oncocytic and Chromophobe Renal Tumours Based on the 5 < sup > th < /sup > Who Classification 2022.

The classification of renal tumors is expanding with the addition of new molecular entities in the 5th World Health Organization classification. Apart from this, the major updates in the definition of papillary renal cell carcinoma are that these tumors are no longer subtyped into type 1 and type 2. In oncocytic tumors, the new molecularly defined renal tumors, emerging and novel entities need to be considered in the diagnosis of oncocytic and chromophobe renal tumors. This is a retrospective study to review and reclassify papillary, oncocytic, and chromophobe renal tumors based on the new WHO classification and correlate with clinical data, gross, microscopic features, and immunohistochemistry markers. A total of thirteen cases were reviewed and the tumor grade was changed for three out of four cases of papillary renal cell carcinoma and a single case was recategorized and graded. In nine cases of oncocytic and chromophobe renal tumors, the diagnoses were modified in 3 cases. Newly defined molecular renal tumors require advanced immunohistochemistry markers and molecular tests. This poses diagnostic challenges to pathologists practicing in low resource settings where molecular tests are not available.

The World Health Organization 2022 Classification of renal tumors: key updates for urologists.

The World Health Organization 2022 Classification of renal tumors: key updates for urologists.

Establishment and Characterization of a TFE3-rearranged Renal Cell Carcinoma Cell Line (FU-UR-2) With the PRCC-TFE3 Fusion Transcript.

Xp11.2-RCC was classified into molecularly defined renal carcinomas and named TFE3-rearranged renal cell carcinoma (TFE3-rRCC) in the 2022 World Health Organization classification of renal tumors. In this study, we established and characterized a TFE3-rRCC cell line from a right-sided renal tumor of a 35-year-old female patient and named it FU-UR-2. FU-UR-2 had been initially diagnosed as a papillary RCC because the patient was 35 years old, a routine immunohistochemical staining for TFE3 was negative, and its morphology was papillary. The G-band analysis revealed an X-chromosome aberration, thus we performed immunohistochemical re-staining for TFE3 and examined the aberration in the TFE3 gene by reverse-transcriptase polymerase chain reaction and fluorescence in situ hybridization. FU-UR-2 was confirmed as a TFE3-rRCC with a PRCC-TFE3 fusion transcript. Cultured FU-UR-2 cells continuously propagated over 90 passages and may provide a new permanent culture model to study pathogenetic mechanisms, investigate biological behavior, and develop new treatments such as molecular-targeting antitumor agents or immunological drugs for TFE3-rRCCs.

Classification of Benign and Malignant Renal Tumors Based on CT Scans and Clinical Data Using Machine Learning Methods

Up to 20% of renal masses ≤4 cm is found to be benign at the time of surgical excision, raising concern for overtreatment. However, the risk of malignancy is currently unable to be accurately predicted prior to surgery using imaging alone. The objective of this study is to propose a machine learning (ML) framework for pre-operative renal tumor classification using readily available clinical and CT imaging data. We tested both traditional ML methods (i.e., XGBoost, random forest (RF)) and deep learning (DL) methods (i.e., multilayer perceptron (MLP), 3D convolutional neural network (3DCNN)) to build the classification model. We discovered that the combination of clinical and radiomics features produced the best results (i.e., AUC [95% CI] of 0.719 [0.712–0.726], a precision [95% CI] of 0.976 [0.975–0.978], a recall [95% CI] of 0.683 [0.675–0.691], and a specificity [95% CI] of 0.827 [0.817–0.837]). Our analysis revealed that employing ML models with CT scans and clinical data holds promise for classifying the risk of renal malignancy. Future work should focus on externally validating the proposed model and features to better support clinical decision-making in renal cancer diagnosis.

Treatment of focal anaplastic Wilms tumor (FAWT): A report from the Children’s Oncology Group (COG) AREN0321 and AREN03B2 studies.

10005 Background: FAWT represents tumors with circumscribed small areas of anaplastic nuclear changes confined to the kidney within an otherwise favorable histology WT. In National Wilms Tumor Study-5, patients with stage I FAWT treated with vincristine and dactinomycin (VA) without flank radiation had 4-year event-free survival (EFS) and overall survival (OS) of 67.5% and 88.9%, respectively. Patients with stage II-IV FAWT were treated with VA plus doxorubicin and radiation (DD4A). Of the 20 patients with stage II-III FAWT, 4 had a relapse and 3 died. The 4-year EFS and OS for stage IV FAWT/upfront nephrectomy were 61.4% and 71.6%, respectively. The COG AREN0321 study evaluated whether intensifying therapy for stage I and IV FAWT would improve survival of these patients. Methods: Tumor histology and stage were confirmed for all patients by real-time central pathology, surgery, and radiology review on protocol AREN03B2, the COG renal tumor classification, biology, and banking study. Patients then enrolled on AREN0321 with stage I-III FAWT received DD4A and those with stage IV FAWT received vincristine, doxorubicin, cyclophosphamide, carboplatin and etoposide with radiation (UH-1). We analyzed outcomes of patients with FAWT treated on AREN0321 and those enrolled only on AREN03B2 who received the same chemotherapy regimens. Results: A total of 46 patients (25 AREN0321 and 21 AREN03B2) were included in the analyses. Of the 25 AREN0321 patients, 17 (68%) had upfront nephrectomy and all 25 received radiation. Of the 5 AREN0321 patients with stage IV FAWT, 4 had metastasis to lung and 1 to lung and liver. Of the 21 AREN03B2 patients, 19 (90%) had upfront nephrectomy and 17 (81%) reported receiving radiation. The 2 AREN03B2 stage IV patients had lung only metastases. The 4-year EFS and OS are displayed in the table below. No events occurred in stage I-II AREN0321 patients nor in any AREN03B2 patients. Of 10 patients with stage III FAWT on AREN0321, 3 had a relapse. None of the 7 stage IV FAWT patients who received UH-1 had a relapse, but one patient died due to toxicity. Conclusions: Patients with stage I and II FAWT had outstanding survival when treated with DD4A and radiation. Intensification of therapy may be warranted for stage III FAWT and appears to have improved patient survival for stage IV FAWT but with increased risk of toxicity. Strategies to alleviate toxicity of UH-1 therapy have been incorporated into the ongoing COG trial for high-risk patients. Clinical trial information: NCT00335556 . [Table: see text]

Eosinophilic solid and cystic renal cell carcinoma with TSC2 mutation: a case report and literature review

BackgroundEosinophilic solid and cystic renal cell carcinoma (ESC-RCC) is an under-recognized, emerging new entity of sporadic renal neoplasms, which is listed as a rare type of renal cell carcinoma in the 2022 WHO renal tumor classification. It is easily misdiagnosed because its characteristics are insufficiently understood.Case presentationWe report one case of ESC-RCC, a 53-year-old female patient with a right kidney mass found during clinical examination. The patient experienced no discomforting symptoms. Computer-tomography imaging at our urinary department showed a round soft tissue density shadow around the right kidney. Microscopic examination revealed a tumor displaying a solid-cystic composition of eosinophilic cells with unique features, revealed by characteristic immunohistochemical markers (CK20-positive/CK7-negative), and a nonsense mutation in TSC2. Ten months after the renal tumor resection, the patient presented in good condition with no recurrence or metastasis.ConclusionsThe distinct morphological, immunophenotypic, and molecular characteristics of ESC-RCC we describe here, based on our case and the relevant literature, highlight the key points of the pathological and differential diagnosis of this novel renal neoplasm. Our findings will therefore deepen our understanding of this novel renal neoplasm and help reduce misdiagnosis.

Update on classification of oncocytic neoplasms of the kidney.

This review provides a summary of recent developments in classification of renal oncocytic neoplasms that were incorporated in the fifth edition WHO classification of renal tumors, released in 2022. Besides the distinct entities of renal oncocytoma and chromophobe renal cell carcinoma, the WHO now acknowledges a heterogeneous group of oncocytic tumors of the kidney that can be reported as 'oncocytic renal neoplasms of low malignant potential'. Case series by multiple institutions have revealed recurrent patterns of morphological features, protein marker expression, and genetic alterations within these neoplasms that may permit further subclassification in the future. The new classification system provides pathologists with the opportunity to simplify the diagnostic workup and reporting of morphologically equivocal oncocytic neoplasms.

Cystic Renal Cell Carcinoma – Sometimes A Diagnostic Dilemma: A Case Report on Rare Multilocular Cystic Renal Cell Carcinoma

Cystic renal tumors often create diagnostic dilemma. It might be a cancer in a complex cyst, cystic degeneration of a renal cell carcinoma, Papillary cystic renal cell carcinoma or multilocular cystic renal cell carcinoma (MCRCC). MCRCC is a rare cystic tumor of the kidney. It is recognized as a separate subtype of renal cell carcinoma in WHO classification of adult renal tumors in 2004 with excellent prognosis. A 47 years old male presented with an incidentally detected large right renal cystic lesion in the upper and hilar region. After discussion with the patient about the possibilities, right radical nephrectomy was performed. Histopathology was done by one of the renowned Pathologists of the country and was reported as a rare variety Multilocular Cystic Renal Cell Carcinoma. It is very important to differentiate this type of cystic renal cell carcinoma from cystic degeneration of conventional RCC. Bangladesh J. Urol. 2021; 24(2): 224-228

Renal oncocytoma: a challenging diagnosis.

The aim of the review is to provide an update on the current and evolving approaches to diagnosing the challenging clinical entity of renal oncocytoma. Renal oncocytoma is being increasingly recognized among patients with renal masses, and it can be found in up to 50% of benign small renal masses (SRMs) less than 4 cm. Renal oncocytomas have benign clinical biology but distinguishing them from some of the other renal masses with more malignant potential can be challenging due to overlapping imaging, histologic, and immunophenotypic characteristics. Increasing integration of various imaging modalities, histologic characteristics, cytogenetics, and molecular and metabolic signatures is helping better define and characterize renal masses. Evolving and complementary diagnostic approaches, including at the molecular level, are continuing to help refine the classification of renal tumors, with implications on their clinical behavior and ultimately clinical management.

Histopathological pattern of adult renal tumours in a tertiary cancer center

Background: Renal tumours are 16th most common malignancies in the world accounting for 2.2% of all new malignancies and 1.8% of all cancer deaths. Histopathological classification of renal tumours along with tumour grade and stage is proven to have prognostic value. Material and Methods: This was a retrospective study conducted in Department of Pathology at B.P. Koirala Memorial Cancer Hospital Bharatpur from January 2017 to December 2019 including cases of partial and radical nephrectomies of patients more than sixteen years of age. Results: Total 78 nephrectomy specimens were received out of which 70 (89.7% consisted of malignant tumour and 8 (10.3%) were benign lesions. Clear cell type renal cell carcinoma was the most common malignant tumour consisting of 57 (81.4%) cases of renal cell carcinoma, followed by papillary type 9(12.9%), Cchromophobe (4.3%) and one case (1.3%) of multiloculated cystic renal neoplasm was noted. Renal cell carcinoma was seen in age range of 26 to 80 years with majority of patients more than 50 years of age, while a small peak (14.3%) was noted in female patients of 31 to 40 years of age. Conclusions: This study found increased incidence of renal cell carcinoma among female patients of 31 to 40 years of age in contrast to studies done in the Western population.

Improving segmentation and classification of renal tumors in small sample 3D CT images using transfer learning with convolutional neural networks.

Computed tomography (CT) images can display internal organs of patients and are particularly suitable for preoperative surgical diagnoses. The increasing demands for computer-aided systems in recent years have facilitated the development of many automated algorithms, especially deep convolutional neural networks, to segment organs and tumors or identify diseases from CT images. However, performances of some systems are highly affected by the amount of training data, while the sizes of medical image data sets, especially three-dimensional (3D) data sets, are usually small. This condition limits the application of deep learning. In this study, given a practical clinical data set that has 3D CT images of 20 patients with renal carcinoma, we designed a pipeline employing transfer learning to alleviate the detrimental effect of the small sample size. A dual-channel fine segmentation network (FS-Net) was constructed to segment kidney and tumor regions, with 210 publicly available 3D images from a competition employed during the training phase. We also built discriminative classifiers to classify the benign and malignant tumors based on the segmented regions, where both handcrafted and deep features were tested. Our experimental results showed that the Dice values of segmented kidney and tumor regions were 0.9662 and 0.7685, respectively, which were better than those of state-of-the-art methods. The classification model using radiomics features can classify most of the tumors correctly. The designed FS-Net was demonstrated to be more effective than simply fine-tuning on the practical small size data set given that the model can borrow knowledge from large auxiliary data without diluting the signal in primary data. For the small data set, radiomics features outperformed deep features in the classification of benign and malignant tumors. This work highlights the importance of architecture design in transfer learning, and the proposed pipeline is anticipated to provide a reference and inspiration for small data analysis.

Unclassified Oncocytic Renal Neoplasm with Rib and Liver Metastases: Metastatic Oncocytoma?

Renal tumors with oncocytic or chromophobe-like morphology can be a common source of diagnostic difficulty. In some series, they constitute the largest group of unclassified renal cell carcinomas, a term used for neoplasms that do not fit the current classification of renal tumors. We describe the histological, immunohistochemical, and molecular findings of an eosinophilic renal neoplasm which presented with rib and liver metastases, and provide a review of the literature. The possibility of a renal oncocytoma with metastases was initially considered but excluded on the basis of several morphological and immunohistochemical features. Additionally, the tumor did not correspond with other traditional or newly emerging categories of renal neoplasms. It was therefore regarded as an unclassified oncocytic renal neoplasm which demonstrated evidence of malignant potential due to the presence of multiple metastases.