No reliable parameters exist that can consistently predict the tumor subtype and associated clinical course upon initial radiographic detection of a mass in the central nervous system.  The analysis of large radiology data sets (radiomics) offers opportunities for non-invasive characterization and tracking of tumors.

Dr. Bi’s team applies artificial intelligence algorithms, such as deep learning, to images, to recognize complex patterns, with strong association to the biological characteristics of the tumor and clinical outcome. Initial results have demonstrated the association of tumor phenotype, based on its imaging appearance, with tumor grade and specific genetic features.  Dr. Bi’s work is focusing on  tracking and predicting tumor behavior over time and improve our ability to inform management decisions and patient counseling.

Dr.Bi’s Publications in PubMed

Dr. Cho is the recipient of the Sperling Family Fellowship 2018.

Advanced brain cancers, including glioblastoma, are extremely deadly and incurable. Even after undergoing surgery followed by chemotherapy, radiotherapy or both, patients life expectancy is limited because:

• Aggressive brain cancer cells tend to invade and incorporate themselves into the surrounding normal brain tissue.
• Chemotherapy drugs are often unable to reach these cells, as they are protected by a highly-evolved barrier, called the blood-brain-barrier (BBB) that separates brain tissues from circulating blood. 
• Chemotherapy is very toxic to normal tissues and can cause severe side effects.

Dr. Cho has established a collaboration among Brigham and Women’s Hospital, the Broad Institute and Massachusetts Institute of Technology, aimed at developing a high-throughput platform that can be used to accurately model drug delivery across the blood brain barrier to reach the tumor cells. Additionally, she is also working to develop new therapeutic drugs which can recognize and kill tumor cells without harming healthy tissues.

Dr. Cho’s Publications in PubMed

In the era of widespread mammography, patients are frequently diagnosed with “high-risk lesion” (HRL) of the breast, a group of abnormalities that confer an increased risk of future breast cancer. The most common HRL is called Atypical Ductal Hyperplasia (ADH), usually treated with surgical excision to ensure that breast cancer is not present in the adjacent tissue. Each year nearly 80,000 women will undergo surgery unnecessarily due to the lack of diagnostic tools to precisely detect ADH.

Dr. King and her team utilize machine learning to identify the subset of patients diagnosed with ADH who are most likely to be upgraded to cancer. Their hypothesis is that machine learning algorithms applied to digitized mammographic and pathologic images, combined with clinical data, can provide a novel and much needed strategy to identify those women who require surgical excision and those that don’t sparing a large percentage of women an unnecessary operation

Dr.King’s Publications in PubMed

Cancer cells are driven by genetic alterations. A particular group of genetic alterations that may cause cancer are called gene fusions, and result from the rearrangement of chromosomal material, which brings together fragments of distant, previously unconnected genes, to create novel genetic sequences. Oncogenic gene fusions are abnormal and are generally present in every cancer cell in a given tumor, and only in the cancer cells in a given patient. For these reasons, gene fusions are extremely useful biomarkers to accurately classify tumors and constitute ideal therapeutic targets to selective attack the cancer cells. Gene fusions are very frequent in some tumor types, such as sarcoma, a type of cancer that is particularly difficult to diagnose and to treat. At present, the identification of sarcoma gene fusions in clinical practice remains challenging and requires complicated methods which are time-consuming and expensive.

Dr. Marino-Enriquez’s team apply the latest Next Generation Sequencing technologies to detect oncogenic gene fusion transcripts in patients’ tumor samples. The goal is to design and implement a reliable, cost-effective, and clinically useful gene fusion detection assay.

Dr. Marino-Enriquez’s Publications in PubMed

Endometrial cancer is the most common gynecologic malignancy, with approximately 40,000 new cases and 7,400 deaths per year in the US. The diagnosis of EIN, the precursor lesion that gives rise to endometrial cancer, is challenging for pathologists but the diagnosis has major treatment implications because EIN is treated with hysterectomy while a benign diagnosis is not.

Computational tools have the potential to standardize diagnostic practice by reigning in this lack of agreement.

1. Before rendering an EIN diagnosis, the pathologist needs to localize the diagnostic area within a slide, which can be a challenge when a lesion is small relative to the amount of tissue being examined. Dr. Papke is interested in developing software that highlights a region likely to contain precancer to draw the pathologist’s attention to the area and ensure that the diagnosis is not missed.
2. The problem of localizing and classifying glandular lesions is not limited to the endometrium; it is in fact the most common diagnostic question in daily practice, including in biopsies and resections of breast, prostate, lung and the GI tract, among other systems. Dr. Papke is interested in generalizing the approach taken in the endometrium to address diagnostic challenges in these other systems as well.

Dr. Papke’s publications in Pubmed

Dr. Tempany leads the The National Center for Image Guided Therapy (NCIGT), located at Brigham and Women’s Hospital and Harvard Medical School in Boston. NCIGT is a Biomedical Technology Resource Center funded by the National Institute of Biomedical Imaging and Bioengineering which serves as a national resource for all aspects of research into medical procedures enhanced by imaging, with the common goal of providing more effective patient care.

Dr. Tempany is particulary involved in the Prostate Project. There are two complex issues that drive the clinical need to change current paradigms for prostate cancer (PCa): The inability to predict aggressiveness of a given cancer, which in turn leads to over treatment, and the increasing evidence that disease progression in men with seemingly low-risk PCa is due to inadequate biopsy sampling. Technical solutions to address these challenges, and their validation in clinic, are lacking.

1. We are working to address these challenges by integrating innovative MR image acquisition and analysis with the MR targeted biopsy platform.
2. We are developing a diagnostic biomedical imaging platform to detect, characterize and diagnose prostate cancer and will provide new opportunities to understand the aggressiveness and heterogeneity of prostate cancer and ultimately allow for development and testing of new predictive markers in focal therapy.

Dr. Tempany’s Publications in PubMed