A Texture Based Pattern Recognition Approach to Distinguish Melanoma from Non-Melanoma Cells in Histopathological Tissue Microarray Sections

Rexhepaj, Elton; Agnarsdóttir, Margrét; Bergman, Julia; Edqvist, Per‐Henrik; Bergqvist, Michael; Uhlén, Mathias; Gallagher, William M.; Lundberg, Emma; Pontén, Fredrik

doi:10.1371/journal.pone.0062070

Cited by 21 publications

(28 citation statements)

References 39 publications

(43 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…[2][3][4][5][6][7] This technology can become fundamental in exploring the human proteome considering the information acquired from human genome efforts [17][18][19] since it can provide a means to explore, in a rapid and costeffective manner, the huge number of antibodies being generated from the non-redundant protein database. In addition, both proteomic and genomic data can be simultaneously obtained from the same specimen.…”

Section: Discussionmentioning

confidence: 99%

Evaluation of Cross-Sample Contamination in Tissue Microarrays by Polymerase Chain Reaction

D'Urso

Spada

Tramonte

et al. 2015

Biopreservation and Biobanking

View full text Add to dashboard Cite

In the past decade, the popularity and power of Tissue Microarray (TMA) technology has increased since it provides a method to detect diagnostic and prognostic markers in an array of clinical tissue specimens collected for translational research. TMAs allow for rapid and cost-effective analysis of hundreds of molecular markers at the nucleic acid and protein levels. This technology is particularly useful in the realization of the Human Protein Atlas Project, since it aims to create a reference database of non-redundant human proteins. In this context, it is important to assure the lack of cross-sample contamination due to the repeated use of the same needle in consecutive coring. Here we show that carry-over contamination from one tissue core to another does not occur, reinforcing the accuracy of the TMA technology in the simultaneous testing of multiple bio-samples.

show abstract

Section: Discussionmentioning

confidence: 99%

Evaluation of Cross-Sample Contamination in Tissue Microarrays by Polymerase Chain Reaction

D'Urso

Spada

Tramonte

et al. 2015

Biopreservation and Biobanking

View full text Add to dashboard Cite

show abstract

“…As a result, C-Path has identified that features of the stromal tissue adjacent to the cancer were better predictors of patient survival than the features of tumor cells alone. Similar quantitative algorithms have been developed for the automated quantification of IHC-determined protein expressions in melanocytes (27), for the automated assessment of the extent of malignant nuclei in colon cancer histology images (28), and for quantifying the architectural complexity in breast and prostate cancer specimens (29, 30). Our group also used quantitative segmentation algorithms to automatically determine scores for HER2-positive, ER-positive, and PR-positive cells within digitized histology images of breast cancer tissues (31), to correlate patient ER statuses with spatial distributions of ER-positive and ER-negative cells, as well as tumor vascular density and vascular area (32), and to determine the extent of the drug-mediated cytotoxic and apoptotic effects in osteosarcoma xenotransplants (33).…”

Section: Evaluation Of the Hypothesismentioning

confidence: 99%

Diagnostic assessment of osteosarcoma chemoresistance based on Virtual Clinical Trials

2015

View full text Add to dashboard Cite

Osteosarcoma is the most common primary bone tumor in pediatric and young adult patients. Successful treatment of osteosarcomas requires a combination of surgical resection and systemic chemotherapy, both neoadjuvant (prior to surgery) and adjuvant (after surgery). The degree of necrosis following neoadjuvant chemotherapy correlates with the subsequent probability of disease-free survival. Tumors with less than 10% of viable cells after treatment represent patients with a more favorable prognosis. However, being able to predict early, such as at the time of the pre-treatment tumor biopsy, how the patient will respond to the standard chemotherapy would provide an opportunity for more personalized patient care. Patients with unfavorable predictions could be studied in a protocol, rather than a standard setting, towards improving therapeutic success. The onset of necrotic cells in osteosarcomas treated with chemotherapeutic agents is a measure of tumor sensitivity to the drugs. We hypothesize that the remaining viable cells, i.e., cells that have not responded to the treatment, are chemoresistant, and that the pathological characteristics of these chemoresistant tumor cells within the osteosarcoma pre-treatment biopsy can predict tumor response to the standard-of-care chemotherapeutic treatment. This hypothesis can be tested by comparing patient histopathology samples before, as well as after treatment to identify both morphological and immunochemical cellular features that are characteristic of chemoresistant cells, i.e., cells that survived treatment. Consequently, using computational simulations of dynamic changes in tumor pathology under the simulated standard of care chemotherapeutic treatment, one can couple the pre- and post-treatment morphological and spatial patterns of chemoresistant cells, and correlate them with patient clinical diagnoses. This procedure, that we named ‘Virtual Clinical Trials’, can serve as a potential predictive biomarker providing a novel value-added decision support tool for oncologists.

show abstract

“…As a result, C-Path has identified that features of the stromal tissue adjacent to the cancer were better predictors of patient survival than the features of tumor cells alone. Similar quantitative algorithms have been developed for the automated quantification of IHC-determined protein expressions in melanocytes (27), for the automated assessment of the extent of malignant nuclei in colon cancer histology images (28), and for quantifying the architectural complexity in breast and prostate cancer specimens (29,30). Our group also used quantitative segmentation algorithms to automatically determine scores for HER2-positive, ERpositive, and PR-positive cells within digitized histology images of breast cancer tissues (31), to correlate patient ER statuses with spatial distributions of ER-positive and ER-negative cells, as well as tumor vascular density and vascular area (32), and to determine the extent of the drug-mediated cytotoxic and apoptotic effects in osteosarcoma xenotransplants (33).…”

Section: Evaluation Of the Hypothesismentioning

confidence: 99%

Diagnostic Assessment of Osteosarcoma Chemoresistance Based on Virtual Clinical Trials

Rejniak

Lloyd

Reed

2015

Preprint

View full text Add to dashboard Cite

Osteosarcoma is the most common primary bone tumor in pediatric and young adult patients. Successful treatment of osteosarcomas requires a combination of surgical resection and systemic chemotherapy, both neoadjuvant (prior to surgery) and adjuvant (after surgery). The degree of necrosis following neoadjuvant chemotherapy correlates with the subsequent probability of disease-free survival. Tumors with less than 10% of viable cells after treatment represent patients with a more favorable prognosis. However, being able to predict early, such as at the time of the pre-treatment tumor biopsy, how the patient will respond to the standard chemotherapy would provide an opportunity for more personalized patient care. Patients with unfavorable predictions could be studied in a protocol, rather than a standard setting, towards improving therapeutic success. The onset of necrotic cells in osteosarcomas treated with chemotherapeutic agents is a measure of tumor sensitivity to the drugs. We hypothesize that the remaining viable cells, i.e., cells that have not responded to the treatment, are chemoresistant, and that the pathological characteristics of these chemoresistant tumor cells within the osteosarcoma pre-treatment biopsy can predict tumor response to the standard-of-care chemotherapeutic treatment. This hypothesis can be tested by comparing patient histopathology samples before, as well as after treatment to identify both morphological and immunochemical cellular features that are characteristic of chemoresistant cells, i.e., cells that survived treatment. Consequently, using computational simulations of dynamic changes in tumor pathology under the simulated standard of care chemotherapeutic treatment, one can couple the pre-and posttreatment morphological and spatial patterns of chemoresistant cells, and correlate them with patient clinical diagnoses. This procedure, that we named 'Virtual Clinical Trials', can serve as a potential predictive biomarker providing a novel value added decision support tool for oncologists.

show abstract

A Texture Based Pattern Recognition Approach to Distinguish Melanoma from Non-Melanoma Cells in Histopathological Tissue Microarray Sections

Cited by 21 publications

References 39 publications

Evaluation of Cross-Sample Contamination in Tissue Microarrays by Polymerase Chain Reaction

Evaluation of Cross-Sample Contamination in Tissue Microarrays by Polymerase Chain Reaction

Diagnostic assessment of osteosarcoma chemoresistance based on Virtual Clinical Trials

Diagnostic Assessment of Osteosarcoma Chemoresistance Based on Virtual Clinical Trials

Contact Info

Product

Resources

About