Farkas Sükösd

Papillary renal cell carcinoma embedded in an oncocytoma: Case report of a rare combined tumour of the kidney

An asymptomatic 1-cm large papillary renal cell carcinoma (RCC) embedded in a 3.5-cm large oncocytoma was diagnosed and removed by right nephrectomy in a 68-year-old male investigated for the abdominal symptoms associated with cholelithiasis. The papillary RCC displayed positive immunohistochemical stainings with cytokeratin 7, alpha-methylacyl-CoA racemase and vimentin and was negative for the E-cadherin and CD117 immunostains, whereas the oncocytoma part showed opposite staining patterns. No gains of chromosomes 7 and 17 or loss of chromosome Y was detected in the papillary carcinoma by fluorescent in situ hybridization with centromeric enumeration probes. This finding is in keeping with the morphologic diagnosis of type 2 papillary RCC reported to have lower rates of these characteristic chromosomal changes. The combination of papillary RCC and oncocytoma, two tumours of different postulated origin, is extremely rare. It may represent a simple coincidence, but 2 previous cases and our current one share a few features, including the intimate embedment of the papillary RCC in the oncocytoma, the small size of the RCC and the old age of the patients. This case raises the point that renal oncocytomas can contain a hidden malignant tumour.

Intelligent image-based in situ single-cell isolation

Quantifying heterogeneities within cell populations is important for many fields including cancer research and neurobiology; however, techniques to isolate individual cells are limited. Here, we describe a high-throughput, non-disruptive, and cost-effective isolation method that is capable of capturing individually targeted cells using widely available techniques. Using high-resolution microscopy, laser microcapture microscopy, image analysis, and machine learning, our technology enables scalable molecular genetic analysis of single cells, targetable by morphology or location within the sample.The isolation of single cells while retaining context is important for quantifying cellular heterogeneity but technically challenging. Here, the authors develop a high-throughput, scalable workflow for microscopy-based single cell isolation using machine-learning, high-throughput microscopy and laser capture microdissection.

Simultaneous detection of BRCA mutations and large genomic rearrangements in germline DNA and FFPE tumor samples

The development of breast and ovarian cancer is strongly connected to the inactivation of the BRCA1 and BRCA2 genes by different germline and somatic alterations, and their diagnosis has great significance in targeted tumor therapy, since recently approved PARP inhibitors show high efficiency in the treatment of BRCA-deficient tumors. This raises the need for new diagnostic methods that are capable of performing an integrative mutation analysis of the BRCA genes not only from germline DNA but also from formalin-fixed and paraffin-embedded (FFPE) tumor samples. Here we describe the development of such a methodology based on next-generation sequencing and a new bioinformatics software for data analysis. The diagnostic method was initially developed on an Illumina MiSeq NGS platform using germline-mutated stem cell lines and then adapted for the Ion Torrent PGM NGS platform as well. We also investigated the usability of NGS coverage data for the detection of copy number variations and exon deletions as a replacement of the conventional MLPA technique. Finally, we tested the developed workflow on FFPE samples from breast and ovarian cancer patients. Our method meets the sensitivity and specificity requirements for the genetic diagnosis of breast and ovarian cancers both from germline and FFPE samples.

Expression of ERCC1, RRM1, TUBB3 in correlation with apoptosis repressor ARC, DNA mismatch repair proteins and p53 in liver metastasis of colorectal cancer

Liver metastasis in colorectal cancer is common and the primary treatment is chemotherapy. To date, there is no routinely used test in clinical practice to predict the effectiveness of conventional chemotherapy. Therefore, biomarkers with predictive value for conventional chemotherapy would be of considerable benefit in treatment planning. We analysed three proteins [excision repair cross-complementing 1 (ERCC1), ribonucleoside-diphosphate reductase 1 (RRM1) and class III β-tubulin (TUBB3)] in colorectal cancer liver metastasis. We used tissue microarray slides with 101 liver metastasis samples, stained for ERCC1, RRM1 and TUBB3 and established scoring systems (fitted for tissue microarray) for each protein. In statistical analysis, we compared the expression of ERCC1, RRM1 and TUBB3 to mismatch proteins (MLH1, MSH2, MSH6 and PMS2), p53 and to apoptosis repressor protein (ARC). Statistically significant correlations were found between ERCC1, TUBB3 and MLH1, MSH2 and RRM1 and MSH2, MSH6. Noteworthy, our analysis revealed a strong significant correlation between cytoplasmic ARC expression and RRM1, TUBB3 (p=0.000 and p=0.001, respectively), implying an additional role of TUBB3 and RRM1 not only in therapy resistance, but also in the apoptotic machinery. Our data strengthens the importance of ERCC1, TUBB3 and RRM1 in the prediction of chemotherapy effectiveness and suggest new functional connections in DNA repair, microtubule network and apoptotic signaling (i.e. ARC protein). In conclusion, we showed the importance and need of predictive biomarkers in metastasized colorectal cancer and pointed out the relevance not only of single predictive markers but also of their interactions with other known and newly explored relations between different signaling pathways.

Metachronous clear cell carcinoma of the tongue and kidney: a diagnostically challenging coincidence

Clear cell carcinomas (CCCs) account for 1% of carcinomas of the salivary glands. A 63-year-old woman presented with a painless, nonulcerated, nodular mass on the right side of the tongue, without palpable neck nodes. After excision and cryotherapy of the mass, the histologic evaluation revealed CCC. At the age of 55, she had undergone radical nephrectomy for CCC of the kidney which extended into the renal vein (pT3aN0). Although she had remained metastasis-free during the follow-up, the clear cell morphology raised the possibility of late lingual metastasis of the renal CCC. A clinical search for metastases, and a series of immunostainings and analysis of the von Hippel-Lindau gene were therefore performed on paraffin-embedded blocks of both tumors: Primary metachronous CCC of the tongue was diagnosed. This case illustrates the diagnostic challenge posed by CCC of the tongue if there is a history of CCC of the kidney.

Environmental properties of cells improve machine learning-based phenotype recognition accuracy

To answer major questions of cell biology, it is often essential to understand the complex phenotypic composition of cellular systems precisely. Modern automated microscopes produce vast amounts of images routinely, making manual analysis nearly impossible. Due to their efficiency, machine learning-based analysis software have become essential tools to perform single-cell-level phenotypic analysis of large imaging datasets. However, an important limitation of such methods is that they do not use the information gained from the cellular micro- and macroenvironment: the algorithmic decision is based solely on the local properties of the cell of interest. Here, we present how various features from the surrounding environment contribute to identifying a cell and how such additional information can improve single-cell-level phenotypic image analysis. The proposed methodology was tested for different sizes of Euclidean and nearest neighbour-based cellular environments both on tissue sections and cell cultures. Our experimental data verify that the surrounding area of a cell largely determines its entity. This effect was found to be especially strong for established tissues, while it was somewhat weaker in the case of cell cultures. Our analysis shows that combining local cellular features with the properties of the cell’s neighbourhood significantly improves the accuracy of machine learning-based phenotyping.

Loss of CDX2 gene expression is associated with DNA repair proteins and is a crucial member of the Wnt signaling pathway in liver metastasis of colorectal cancer

Caudal type homeobox 2 (CDX2) has been well-established as a diagnostic marker for colorectal cancer (CRC); however, less is known about its regulation, particularly its potential interactions with the DNA repair proteins, adenomatous polyposis coli (APC) and β-catenin, in a non-transcriptional manner. In the present study, the protein expression of CDX2 was analyzed, depending on the expression of the DNA repair proteins, mismatch repair (MMR), O6-methylguanine DNA methyltransferase (MGMT) and excision repair cross-complementing 1 (ERCC1), and its importance in Wnt signaling was also determined. A total of 101 liver metastases were punched into tissue microarray (TMA) blocks and serial sections were cut for immunohistochemistry. For each protein, an immunoreactive score was generated according to literature data and the scores were fitted to TMA. Subsequently, statistical analysis was performed to compare the levels of expression with each other and with clinical data. CDX2 loss of expression was observed in 38.5% of the CRC liver metastasis cases. A statistically significant association between CDX2 and each of the investigated MMRs was observed: MutL Homolog 1 (P<0.01), MutS protein Homolog (MSH) 2 (P<0.01), MSH6 (P<0.01), and postmeiotic segregation increased 2 (P=0.040). Furthermore, loss of MGMT and ERCC1 was also associated with CDX2 loss (P=0.039 and P<0.01, respectively). In addition, CDX2 and ERCC1 were inversely associated with metastatic tumor size (P=0.038 and P=0.027, respectively). Sustained CDX2 expression was associated with a higher expression of cytoplasmic/membranous β-catenin and with nuclear APC expression (P=0.042 and P<0.01, respectively). In conclusion, CDX2 loss of expression was not a rare event in liver metastasis of CRC and the results suggested that CDX2 may be involved in mechanisms resulting in the loss of DNA repair protein expression, and in turn methylation; however, its exact function in this context remains to be elucidated.

Show me your neighbours, and I'll tell you what you are - cellular microenvironment matters

To answer major questions of cell biology, it is essential to understand cellular complexity. Modern automated microscopes produce vast amounts of images routinely, making manual analysis nearly impossible. Due to their efficiency, machine learning-based analysis software have become essential tools to perform single-cell-level phenotypic analysis of large imaging datasets. However, an important limitation of such methods is that they do not use the information gained from the cellular micro- and macroenvironment: the algorithmic decision is based solely on the local properties of the cell of interest. Here, we present how various microenvironmental features contribute to identifying a cell and how such additional information can improve single-cell-level phenotypic image analysis. The proposed methodology was tested for different sizes of Euclidean and nearest neighbour-based cellular environments both on tissue sections and cell cultures. Our experimental data verify that the microenvironment of a cell largely determines its entity. This effect was found to be especially strong for established tissues, while it was somewhat weaker in the case of cell cultures. Our analysis shows that combining local cellular features with the properties of the cells microenvironment significantly improves the accuracy of machine learning-based phenotyping.

What can be more prognostic than the pTNM category assessed on radical cystectomy specimens

Sir, The ideal way to compare methods of macroscopic examination would be to examine the same samples twice independently. However, once a radical cystectomy specimen has been completely embedded, alternative methods cannot be investigated because no material remains. Gaisa et al. elegantly cut this Gordian knot with their “virtual superimposed approach.” They embedded the whole cystectomy specimen into small blocks and then extracted the block data that would have been used for standard assessment on the basis of specimen photographs and compared them. As concerns tumor stage, they did not find any significant differences between the two approaches [1]. Another novelty they introduced is the pTsum category. They determined tumor stage on the cystectomy specimen, but in cases with previous TUR-B and no or minimal residual tumor, the final tumor stage was defined as the sum of the TUR-B and the cystectomy specimen. Theoretically, it is possible that the pTsum category is prognostically more informative than the pT as defined in the standard TNM system. However, in their paper, the Kaplan-Meier curve of overall survival in relation to pTsum (Fig. 4A in the publication) suggests that the death rate of patients with pT1 tumors is higher than that of patients with pT2 tumors after a 20-month follow-up. We consider that introduction of the pTsum category is not in agreement with the authors’ intended study design in terms of comparing the standard specimen sampling with whole specimen embedding relative to standard pTNM. A significant degree of uncertainty is introduced with unknown TUR-B results. How was “minimal residual tumor” defined, which TUR-B stage had more impact than pTNM on the cystectomy specimen, and what is the advantage of having an early TUR-B stage? Why was the traditional staging system not used? It is rather disturbing that the authors wrote about stage but studied stage only with a self-created pTsum system. Since 2008, we have examined 315 cystectomy specimens and used whole-organ embedding approach with commercially available macroblocks as part of the daily routine. We have reported stage distribution of the first 138 cases and compared this with data obtained by standard methods on 15,586 literature cases [2]. Our method highly reliably mirrors the averages of published data, with as only exception that we identified a significantly higher proportion of pT4 cases. When pTsum stages are compared with our data and literature-based data, two differences emerge (Fig. 1). First, pT1–pT3 stages are more frequent in the series by Gaisa et al. than in ours. This may be due to the pTsum category, with TUR-B stages increasing the case number of higher stage cases at the expense of lower stages. Second, Gaisa et al. found nearly the same proportion of pT4 cases as the average we generated from literature data. In contrast, our series reveals a significantly higher rate of this advanced category because our base block approach, in which the prostatebladder boundary is processed in 12 radial cuts, detects prostate involvement more precisely. We therefore regard the whole-organ embedding approach as the gold standard in the assessment of the pTNM category and provide prognostically significant histopathological parameters in radical cystectomy specimens. * Farkas Sükösd sukosd.farkas@med.u-szeged.hu