Department of Pathology

Joining Forces to Validate a Novel Biomarker for Breast Cancer Metastasis


 _Maja_Oktay_headshot   Jones Joan_kk[1]   2952-joseph-sparano   condeelis    
Maja Oktay, MD, PhD Joan Jones, MD Joseph Sparano, MD  John Condeelis, PhD  David Entenberg, MSc


In December 2016, Joseph Sparano, MD, associate chair of medical oncology at the Montefiore Einstein Center for Cancer Care, presented an invited paper at the 39th San Antonio Breast Cancer Symposium in San Antonio, TX.  Dr. Sparano reported on a study he led to clinically validate MetaSite (MS), a novel biomarker for metastasis in breast cancer. Joining him were two of his coauthors, Joan Jones, MD, and Maja Oktay, MD,PhD—both members of the Montefiore/Einstein Department of Pathology and Einstein’s Department of Anatomy & Structural Biology.

Doorway for Metastasis 

Subpopulations of tumor cells can form micro-anatomic associations with endothelial cells and macrophages. These structures, known as “tumor microenvironments of metastasis” (TMEM or MetaSites), are the point of entry for cancer cells into blood vessels and distant spread. First identified in mice, TMEM also occur in human early-stage breast cancers. 


“TMEM are the doorway for metastasis,” says Dr. Sparano. “Tumor size and lymph-node metastases are time-dependent prognostic factors.” The risk of metastasis is a function of these two factors along with treatment sensitivity (Recurrence score) and risk of dissemination (MetaSite or MS score), he explains.

The Montefiore/Einstein study is the third to show an association between MetaSites and  recurrence. The researchers analyzed tumors from 600 patients with early-stage breast cancers who had undergone surgery, adjuvant chemotherapy and endocrine therapy. They used a fully automated digital pathology image analysis algorithm to assign TMEM or MS score. 

 TMEM IHC and Automation Summary Image
Left: identification of TMEM by immunohistochemistry (IHC) where endothelial cells (EC) are visualized in blue, macropahges (M) in brown and Mena-expressing tumor cells (TC) in red.
Middle: image classified by color deconvolution. Right: automated identification of TMEM.

A Productive Partnership 

Dr. Jones developed the assay for detecting TMEM sites in human breast-cancer tissue samples. She trained Dr. Oktay and Dr. Timothy D’Alfonso of Weill Cornell Medicine (another team member) in TMEM interpretation. Together with John Condeelis, PhD, co-director of Einstein’s Gruss-Lipper Biophotonics Center and its Integrated Imaging Program, and David Entenberg, MSc, director of technological development, Dr. Jones automated the TMEM analysis.  

Acting in his capacity as vice chair of the ECOG/ACRIN--a multidisciplinary group that designs and conducts biomarker-driving cancer research involving adults who have or are at risk of developing cancer--and as study chair for the National Cancer Institute’s TAILORx clinical trial, Dr. Sparano directed the acquisition of the ECOG-ACRIN E2197 cohort for the study analysis.

Drs. Jones, Oktay and D’Alfonso worked with Dr. Sparano to ensure that the tissue sections from the paraffin blocks containing valuable patient material were processed properly (sectioning, preservation, shipment to MetaStat laboratory for TMEM staining, choosing the right fields for analysis of TMEM score and overall pathology QC of the workflow).

A Promising Predictor  

The MetaSite test is the result of pre-clinical work led by Dr. Condeelis, who is also co-chair of anatomy & structural biology at Einstein. Using intravital imaging, Dr. Condeelis’s team identified sites of breast cancer cell intravasation in mouse models to validate TMEM sites. These sites were then identified in formalin-fixed, paraffin-embedded sections of human breast cancers using a triple immunostain approach developed by Dr. Jones.

Dr. Oktay began collaborating with Drs. Condeelis and Jones in 2008, shortly before the first study of the prognostic value of TMEM in breast cancer patients was published. She was involved in all subsequent studies, which showed that TMEM is prognostic of breast cancer metastasis. She also served as lead investigator on a study which showed that invasive breast cancer cells obtained from patients use TMEM sites to intravastate.

Inventing an Algorithm 

Dr. Jones, who directs clinical imaging applications for the IIP, collaborated with David Entenberg, MSc to develop the automated TMEM scoring with assistance from Drs. Oktay, D’Alfonso and Condeelis.

Here’s how it works: Tissue sections are digitized on a whole-slide scanner that acquires thousands of high-resolution images, and stitches them together to form a large mosaicked image of the tissue. The resulting digital pathology image is then processed by custom software which analyzes the staining color of each pixel within the image and assigns it to one of four categories: tumor, macrophages, endothelial cells or stroma. Selection criteria are applied to filter the resulting objects (clusters of pixels) based on their spatial relationships. This identifies the individual TMEM structures. 

While the classification and selection criteria are applied at the individual pixel level, the analysis is conducted across the entire tissue. A final MetaSite score is calculated based on the TMEM count in the top 10 high-resolution images.

The algorithm will be a boon for pathologists. “Our statistical analysis of the algorithm shows that it performs as well as a pathologist in counting TMEM for a MetaSite score, while reducing the pathologist’s time from 50 minutes per case to about 5 minutes,” says Dr. Oktay. The algorithm, she adds, “has a day-to-day reproducibility of 99%.”

Potential to Improve Patient Care  

The MetaSite test can predict early recurrence of metastasis within 5 years. It can benefit breast cancer patients with estrogen-positive tumors not expressing Her2Neu receptor (ER+/ Her2-) by providing prognostic information that complements the information now being collected with clinical and pathological parameters and Oncotype DX score.

A patient’s MS score also can help her doctor in designing her treatment plan. For example, says Dr. Oktay, “We have preliminary data indicating that some patients with low-intermediate Oncotype DX score have high MS scores and subsequently develop distant recurrence. So it would be very helpful to have both types of scores available.”

Further evaluation is needed, says Dr. Sparano, to identify patients at highest risk of recurrence within 5 years who are most likely to benefit from adjuvant chemotherapy or novel therapies.

“This work would not have been possible without a multidisciplinary team,” he adds, “with each member bringing their own unique skills and expertise to the table.”

To learn more, go to: 

Click here to log in